Sie sind auf Seite 1von 18

A New Model of Scheduling in Manufacturing: Tasks, Roles, and Monitoring

Sarah Jackson, John R. Wilson, and Bart L. MacCarthy, University of Nottingham, Nottingham, UK
For over 3 decades there was a belief that computer-based solutions would solve complex industrial scheduling problems, yet most manufacturing organizations still require human contributions for effective scheduling performance. We present a new model of scheduling for the development and implementation of effective scheduling systems within manufacturing companies. The model derives from investigating the work of 7 schedulers in 4 manufacturing environments using a qualitative eld study approach, for which novel eld-based data collection and analysis methods were developed. The results show that scheduling in practice comprises task, role, and monitoring activities and that the business environment inuences a scheduler at work. A new denition of scheduling is presented that includes the signicant facilitation and implementation aspects of human scheduling ignored by many computer-based scheduling approaches. The implications for this model extend across the domains of human factors and operations management, especially for the analysis and improvement of existing and new production planning and control processes and enterprise information systems. Actual or potential applications of this research include the analysis, design, and management of planning, scheduling, and control processes in industry; the selection, training, and support of production schedulers; and the allocation of tasks to humans and computer systems in industrial planning, scheduling, and control processes. INTRODUCTION There are few areas in life in which scheduling does not play a part. Scheduling is a fundamental process that is essential within business, industry, and service sectors. There are numerous forms of scheduling for example, devising personnel rosters, train and aircraft timetables, and administrative timetables, as well as production scheduling. This paper focuses on the domain of production scheduling. In its simplest form, a production schedule is a list or sequence of work that needs to be completed. The objective of the scheduling process is to set the priorities of the work and the allocation of resources to achieve this. A typical denition of scheduling that reects this view is the goal-directed allocation of resources over time to perform a collection of tasks (Baker, 1974; Morton & Pentico, 1993; Ovacik & Uzsoy, 1997; Pinedo, 1995; Sadowski & Medeiros, 1982). Taking this perspective, schedule generation appears to be an optimization and prioritization problem. This classic operations research perspective of the function underpins many previous research solutions for the scheduling problem, building mathematically tractable models using specific algorithms to solve scheduling scenarios. However, the majority of derived mathematical solutions have not been used in practice by manufacturing businesses (MacCarthy & Wilson, 2001). To businesses, such models are static and appear to ignore the complexity of the real world, in which scheduling is both dynamic and ill dened (Stoop & Wiers, 1996).

Address correspondence to John R. Wilson, Institute of Occupational Ergonomics, School of Mechanical, Materials and Manufacturing Engineering, University of Nottingham, Nottingham, NG7 2RD, UK; john.wilson@nottingham.uk. HUMAN FACTORS, Vol. 46, No. 3, Fall 2004, pp. 533550. Copyright 2004, Human Factors and Ergonomics Society. All rights reserved.

534 The sheer volume of published research on classical scheduling theory, as noted by Dessouky, Moray, and Kijowski (1995), demonstrates that scheduling researchers are still searching for models of scheduling that can provide realistic business solutions. Most of this type of research simply adopts the mathematical approach and ignores the significant human contribution to scheduling. However, the study of practice, upon which this paper is based, demonstrates the limitations of this purely computational perspective of scheduling. This paper draws from a program of research that focused on the investigation and implementation of scheduling performance (Crawford, MacCarthy, Wilson, & Vernon, 1999; MacCarthy & Wilson, 2001; MacCarthy, Wilson, & Crawford, 2001). The underlying research assumption was straightforward: Researchers should focus on how scheduling is actually carried out in order to understand what scheduling is and how it occurs in practice. The overall aim was to develop a descriptive model of scheduling performance that would capture scheduler behavior and performance in order to underpin the development of scheduling systems that would be more appropriate than those based on previous theoretical scheduling models. We begin by identifying the need for a new model of scheduling, using evidence from previous research on models of the human scheduler. Our research derives from a naturalistic decision making (NDM) perspective (Klein, Orasanu, Calderwood, & Zsambok, 1993; Lipshitz, Klein, Orasanu, & Salas, 2001). The nature of the study, the research approach, and methods used to conduct the eld studies and to derive the model are rst presented, and this is supported by an overview of the participant schedulers. The main contribution of the paper centers on the development of a new model of scheduling in manufacturing. The implications and potential applications of the model in practice conclude the paper. THE NEED FOR A NEW MODEL OF SCHEDULING Some previous research has exposed the eld of scheduling research to consideration of context and actual working environments; a full review of this field of research is presented in

Fall 2004 Human Factors Crawford and Wiers (2001). Here we concentrate on research that has attempted to develop models of human scheduling in manufacturing contexts. Sanderson (1991) proposed a model human scheduler as a descriptive, computational model that could support the design decisions needed in advanced manufacturing systems and improve understanding of the human contribution to scheduling functions. The model is limited in that it proposes a framework that has not been validated in practice; also, it is built on the assumption that scheduling is a predominantly individual, cognitive, decision making function. As we will show, this is not a safe assumption. Dessouky et al. (1995) developed a model using scheduling theory, with its well-developed quantitative models, as the basis for a systematic conceptual framework of research into behavior in human-machine settings. However, the research is based on limited examples of scheduling behavior in laboratory scenarios and theoretical solutions, rather than on studies of actual practice. Although such theoretical models are an advance over the purely computational operations research models, they are not grounded in studies of practice. We now address studies that are based on empirical studies of schedulers in the eld. McKay, Safayeni, and Buzacott (1995a, 1995b) studied schedulers at work using techniques from the behavioral and cognitive sciences and produced an empirically derived model of the human control necessary within dynamic and uncertain production scheduling situations, considering what scheduling in the real world actually entails. Wiers (1996) extended this work to use a quantitative approach to study the decision behavior of production schedulers using performance, action, and disturbance variables. He concluded that schedulers who schedule comparable production units displayed different types of decision behaviors, but the complexity of the scheduling situations made the ndings difficult to generalize. Wiers (1997) acknowledged that the inability to generalize findings makes it difcult to construct a reliable quantitative model of production scheduling. Higgins (2001) built upon Sandersons (1991) model human scheduler approach in order to develop a methodology for designing scheduling

A NEW MODEL OF SCHEDULING IN MANUFACTURING tools that support schedulers working in dynamic environments. He stated that a schedulers decision making behavior extends beyond the scheduling heuristics perspective of classical scheduling theory. A hybrid intelligent production scheduling system was presented based on a scheduling goal structure model (Higgins, 1999) of scheduling behavior. This model provides a decision making architecture of scheduler behavior generated from field data; however, the work focused mainly on the design of user interfaces, and the model was based on one case study. These studies provided a useful platform for the study reported here. However, we believe that many of the theoretical and empirically derived models are based on a number of invalid assumptions. One frequently implied assumption is that a human schedulers contribution to the scheduling function is mainly cognitive that is, a schedulers job is a series of sequential decisions that lead ultimately to deciding on the sequences of work that in combination produce the schedule. Our early research work and indeed our own industrial experiences support the perspective that there is in fact signicantly more to the scheduling function. A schedulers performance cannot be assessed solely from his or her individual cognitive contribution to the scheduling function; a scheduler also provides signicant social and organizational contributions that support the scheduling function, and the social and organizational setting will provide contextual inuence on scheduling performance. This observation is based on the present authors personal experiences (one has worked as a scheduler in a number of manufacturing businesses), on their research experience, on analysis of empirical studies that have investigated the reality of the scheduling function (see Crawford & Wiers, 2001), and on research within the domains of NDM, cognitive anthropology, and cultural cognition. Work in all of these areas has demonstrated that individual cognitive activity should be considered in context and that context is not a xed set of surrounding conditions but, rather, a wider dynamic process in which cognition of the individual is only one element. Examples of research based on this perspective are presented next. Context-Based Perspectives on the Domain of Scheduling

535

The NDM perspective was a key influence on how we investigated the activities, and specifically the decision making processes, of schedulers in context-based manufacturing situations (see Crawford et al., 1999). NDM research raises questions regarding the ability of classical decision theory to adequately simulate the decision episodes presented by participants working in uncertain, unstable, and rapidly changing environments (for reviews see Einhorn & Hogarth, 1981; Keeney & Raiffa, 1976; Slovic, Fischoff, & Lichenstein, 1977). Classical decision making research tends to focus on the decision event, with a choice made from among a xed set of known alternatives based on stable goals, purposes, and values. In NDM research the focus is on decisions that are embedded in larger dynamic tasks and made by knowledgeable and experienced decision makers, on trying to understand how human decision makers actually make decisions in complex real-world settings, and on learning how to support such processes (Klein et al., 1993; Lipshitz et al., 2001). Inherent uncertainties and other real-world factors are highlighted explicitly in context-based decision making (Orasanu & Connolly, 1993). Such factors were consistent with our previous experience and understanding of the scheduling function. There are other relevant context-based perspectives for example, situated cognition. One well-known work in this field is Cognition in the Wild (Hutchins, 1995), which argues that in order to undertake real-world research, a researcher must move the boundaries of the cognitive unit of analysis beyond the individual to incorporate other personnel and the systems they use in their work, as a combined cognitive and computational system. This approach has many parallels with the NDM perspective. Another related domain is computational organization theory models, which attempt to computationally model and simulate an organization in which information, people, decision responsibility, tasks, resources, and opportunities are distributed geographically, temporally, and structurally (Prietula, Carley, & Gasser, 1998). This approach provides useful insights into how to simulate and model group interactions in terms of organizational

536 behavior. These perspectives influenced how we designed the study reported here and how we subsequently analyzed the data. RESEARCH APPROACH The model presented in this paper emerged from a project investigating human scheduling and related areas within seven manufacturing enterprises. The eld studies were extensive and took an explicitly multi- and interdisciplinary approach. The sponsors in each company were senior planning and scheduling managers; prior to the case study work, a long period was spent with these managers in order to understand the context within which scheduling was carried out and to identify appropriate participants. Details of the participating companies are presented in Table 1. Spatially and temporally distributed activities, such as production scheduling, are necessarily difcult and time consuming to study. The presence of the researchers may be disruptive and inconvenient to some degree. Commercial condentiality must be addressed, as planning and scheduling are at the heart of an organizations real operational systems. Work takes place in a number of different locations, necessitating exible observation methods. Great care must be

Fall 2004 Human Factors taken if data from eld studies are to be of value in supporting strong conclusions. Methodological Approach Following a pilot study it became evident that an appropriate methodology, with effective and efcient data gathering and analysis methods, would be critical to the development of the model. We developed, tested, and used new structured data gathering and analysis methods that were specifically designed to capture dynamic, eld-based data. It was important that data collection took place in the field because only by studying scheduling in its context could the social and organizational influences and effects be properly accounted for. An assessment of the domain literature, the literature on potential research approaches, and the preliminary visits to the participating companies made it evident that a qualitative eld study approach (Miles & Huberman, 1994) would allow a detailed understanding of scheduling practices and the behavior and performance of the human schedulers within a natural environment. Following trials of a large number of methods, the potential set of methods was reduced to a toolkit of six main types of technique: (a) direct observation, (b) retrospective decision probe, (c) structured and semistructured interviews,

TABLE 1: Comparison of the Types of Manufacturing Environments of the Case Study Companies Case Study Data Participant A B C D E F G Industry Sector Textiles Electronic components Aerospace Aerospace Aerospace Steel production Steel production Business Context Discrete manufacturing Non-MRP-driven planning system Discrete manufacturing MRP-driven planning system Discrete manufacturing MRP-driven planning system Discrete manufacturing MRP-driven planning system Discrete manufacturing MRP-driven planning system Continuous process Non-MRP-driven planning system Continuous process Non-MRP-driven planning system No. of Employees <500 500+ 2500 2500 2500 4700 4700

Note. MRP = materials requirements planning.

A NEW MODEL OF SCHEDULING IN MANUFACTURING (d) task analysis, (e) concurrent verbal protocols, and (f) attributes analysis. For more information on all of the methods used in the field study of scheduling, see Crawford et al. (1999) and MacCarthy et al. (2001). The two main methods that underpinned the development of the model presented here were direct observation of the schedulers at work and a retrospective decision probe. These are explained in the following sections. Direct observation. An inductive research approach places great emphasis on the accurate recording and analysis of emergent data, and an important aid to understanding what schedulers do and how they do it was to observe them within their usual scheduling environments. There were practical constraints for instance, video recording was not possible because of location difculties and commercial reasons. Therefore, an observation sheet was developed to allow fast, accurate, and portable data collection and included information on interpersonal interactions, as these had been identied during pilot studies as a critical determinant of scheduler load and strategy. The name or function of the source of the interaction (e.g., sales, production), the type (e.g., face to face, phone, e-mail), and the subject (e.g., query, information transfer) were noted. The apparent effects and consequences were also documented. Retrospective decision probe. Prior to conducting the studies we believed that schedulers make a range of different decisions as part of their job, and the importance of decision mak-

537

ing within the scheduling function has been widely commented on in previous research (see Crawford & Wiers, 2001). We adapted methods used by Milliken (1987) and Lipshitz and Strauss (1997) on uncertainty in decision making for use in the fieldwork conducted in the businesses. Direct structured observation did not allow concurrent questioning about the schedulers decisions; therefore we developed the retrospective decision analysis method to support the ndings of the observation method. In this approach, the researchers note an identification number on the observation sheet when the participant appears to make a decision of any type in carrying out his or her job. At the end of the observation session, participants are retrospectively questioned about each of the noted decisions. A structured interview format, a decision probe, is used so that the same questions are asked for each of the decisions noted. The probe questions are shown in Table 2. The decision probe responses can be represented visually in a decision diagram to collate all of the information about each decision and to present the results to the scheduler for later validation. This collection of responses provides a snapshot of scheduling decision making processes. Each decision diagram also shows the information and knowledge needed by the scheduler to make a decision, the environmental factors that affected the decision, any options that were available to the scheduler, and why they were not chosen. An example of a decision diagram is shown in Figure 1.

TABLE 2: Retrospective Decision Probe Questions Asked of Each Scheduler Subsequent to Daily Direct Observation Session 1. 2. 3. 4. 5. 6. 7. 8. 9. Describe this decision episode in your own words. What caused you to have to make the decision? What was it that you thought about in order to make the decision? What information (as opposed to knowledge) did you use to make this decision? Were you forced to make a decision before you had all the information you needed? What knowledge (as opposed to information) did you use to make this decision? Do you consider this knowledge to be complete? Did you have to communicate the outcome of this decision to anybody else? How would you rate this decision? Complexity: Easy, Average complexity, Complex Time pressure: No time pressure, Some pressure, Extreme pressure Advice or guidance: Never, sometimes, always need it Does a documented procedure exist for this type of problem? Overall, do you consider this decision to be a typical or an untypical type of decision?

10. 11.

538

Figure 1. A decision diagram and associated decision description and decision probe responses of a participant scheduler.

A NEW MODEL OF SCHEDULING IN MANUFACTURING The Participant Schedulers The selection and study of appropriate participant schedulers was an important aspect of the research. The model presented here was proposed on the basis of detailed examination of the work of seven schedulers in four companies. The types of manufacturing processes in which the companies were involved were (a) traditional manufacturing processes in the textile and steel industries and (b) leading-edge hightechnology manufacturing processes in the electronic components and aerospace industries. A summary of the background information gathered about the seven case study schedulers is presented in Table 3. Each scheduler was observed over a 1-week period. Data were written up off line. The researchers returned for a 1-day follow-up visit that enabled clarication of data queries and to validate the data that had been collected. The analysis took place off line, and researchers returned for a nal 2-day visit to present and discuss the results with the schedulers. This research

539

design provided a detailed snapshot of how the schedulers worked within a 1-week period but also built in a longitudinal aspect by enabling the researchers to return on follow-up visits. Data Analysis The data collected from observing and interviewing all of the participant schedulers were analyzed using documented methods of qualitative data analysis (Bernard, 1995; Miles & Huberman, 1994; Robson, 1993; Strauss, 1987). Data analysis was carried out both during data collection and on completion of data collection. The main analysis approach used in the gradual development of the model was coding. Data Coding Codes are labels for assigning units of meaning to the descriptive data compiled during a study (Miles & Huberman, 1994). The data collected using the methods described were coded into two levels: (a) preliminary descriptive codes and (b) second-level pattern coding, which produces explanatory codes that identify emergent

TABLE 3: Comparison of the Education and Work Backgrounds of the Seven Participant Schedulers Scheduler Background Data Participant A B Age 4150 4150 Work Experience Company: 10 years Job: 10 years Company: 24 years Job: 6 months Company: 11 years Job: 4 years Company: 16 years Job: 4 years Company: 19 years Job: 1 year Company: 38 years Job: 35 years Company: 36 years Job: 22 years Training Not known Not known Education High school High school IPCD; IIMC C&G: quality control High school ICPIM High school ICPIM High school C&G: carpentry High school High school ONC: math & engineering C&G: plate, sheet, & strip (metal) Hours of Work Works as required 8 a.m.5 p.m. (day shifts) 8:30 a.m.5:30 p.m. (day shifts) 8:30 a.m.5:30 p.m. (day shifts) 8:30 a.m.5:30 p.m. (day shifts) 8-hour shifts (day & night) 8-hour shifts (day & night)

C D E F G

4150 3140 4150 51+ 51+

None 112 days of MRP training Basic MRP training On-the-job training On-the-job training

Note. IPCD = Institute of Production Control Diploma; IIMC = Institute of Industrial Managers Certicate; C&G = City & Guilds Certicate; ICPIM = Introductory Certicate in Production and Inventory Management; ONC = Ordinary National Certicate.

540 themes (Strauss & Corbin, 1994). To develop preliminary descriptive codes, a summary notation is applied to a chunk of data. An example of the coding process on a section of the observation data is shown in Table 4. This process of applying codes to sections of data continues for all scheduler data, and categories or labels are generated beside each line or paragraph. Once all the data have been coded at the descriptive level, a preliminary master list of codes is drawn up. This is an inductive method in which the researchers code the data in order to see how the data function within the context of the situation and to determine how many categories of codes are appropriate. The analysis produces a code category plus related subcodes. In the example shown in Table 4, the scheduler was checking or looking for information about the situation. This was defined as the assessing the situation (AS) code. Within this master code, the scheduler was sometimes assessing the situation as to the status (ST) of orders (or) or materials (mats). Therefore, for a section of data in which a scheduler was assessing the situation to determine the status of materials, the code AS/ST/mats would be assigned. However, simply classifying phenomena does not constitute a sufcient analysis of the data. It is at this point in the analysis process that any patterns occurring in the data must be understood. Pattern codes are explanatory; they identify an emergent theme and pull together material into meaningful units of analysis. Pattern coding can be thought of as analogous to factor-analytic methods in statistical analysis that attempt to place sets of variables into distinct groups. Miles and Huberman (1994) suggested using

Fall 2004 Human Factors summarizers to build pattern codes. A summarizer can be based on a variety of elements, such as repeatedly observed behaviors, norms, relationships, local meanings and explanations, inferential clusters, and single-case and crosscase elements. An example of building pattern codes using summarizers is shown in Table 5. Pattern codes can be developed by adding tentative codes and trying them out on subsequent data to see if they t. The pattern codes can then be veried and checked during subsequent data collection by all researchers. If the pattern code does not t, it is qualied rather than completely discounted. In this way, pattern coding is a means of data reduction and of funneling ndings for within-case and cross-case analysis. An important aspect of this research was the emphasis placed on recording accurate and reliable data by including feedback and validation sessions into the overall research design. The schedulers made many useful comments within the validation phase for example, explaining the relevance or frequency of a particular type of decision. RESULTS The results relevant to the model, rather than to the larger project, are reported here: first, the types of activities that describe a schedulers behavior and performance within a business, and second, the factors that inuence how schedulers work within a business environment. A critical point is that the activities presented are the results of a very high level of consistency of scheduler behavior and performance: All of the schedulers undertook the key activities described in the next section in some form in each of the business environments. That is, there are coded data and clear patterns demonstrating each of

TABLE 4: Example of the Coding Process on a Section of the Observation Data Scheduler Action (Number 6): Check another order (Number 9): Check parts of valve (Number 15): Scheduler uses information system to see operation status Analyst Code An example of the scheduler checking the status of orders on the shop oor. An example of the scheduler checking the status of materials within the factory. An example of the scheduler checking the status of an order and gathering information from a management information system.

A NEW MODEL OF SCHEDULING IN MANUFACTURING


TABLE 5: Examples of Summarizers and Pattern Codes Used in the Study Summarizer Themes Causes/explanations Relationships among people Emerging constructs Pattern Codes There is no explicit mention of performance measures within the department

541

Releasing goods represents the end of schedulers boundary and responsibility and highlights the schedulers authority within the plant The scheduler is geographically separated/distant from colleagues The scheduler holds the role of a problem predictor and solver within the department

the scheduler activities displayed by each scheduler studied. We believe that these results demonstrate that the ndings are generalizable to other scheduling environments. Scheduler Activities The direct observation method highlighted a number of important issues. First, schedulers work in a variety of ways. Some schedulers sit alone; others sit in an ofce as part of a team of other production personnel. Some schedulers move around their environment, constantly interacting with personnel from other functions and attending meetings. Some schedulers remain sitting at a desk for prolonged periods; others work directly with other business areas to form formal or informal integrated and colocated teams, or they work with personnel located across the business or with suppliers to the business to form more distributed, virtual teams. The observations also demonstrated that schedulers do not often sit down and work through the generation of a schedule in a sequential or structured way. In fact, in the majority of the participating businesses, the schedule is produced by the information system. Therefore, the schedulers main role is not to carry out sequencing or allocation decisions. Instead, the schedulers managed the scheduling function in order to support the transfer of a virtual production plan into production reality. The analysis of the data collected using all of the methods produced pattern codes that demonstrated that the schedulers carried out three types of activities: tasks, roles, and monitoring. Tasks. The term task is used here to describe goal-directed activities that is, activities that have a dened start and end point and an objective. Despite the centrality of the concept of

task in the human factors domain, there is little consensus on the meaning and scope of the term. Kirwan and Ainsworth (1992) dened a task as the method that is adopted to attain a goal that is constrained by a number of contextual factors. Therefore a persons task may be seen as trying to attain a goal in a particular context by utilizing the facilities and resources available while overcoming any constraints that may have been imposed. Stammers and Shepherd (1995) also noted that there are three interacting task components: requirements, environment, and behavior. It is evident, however, that although the basic framework of task activity is determined by context and can therefore be documented by the task requirements and task environment, task behavior can vary greatly between individuals and with experience. In this study, all of the schedulers were subject to formalized goal-directed activities that businesses expect schedulers to undertake, as listed in their job descriptions. An analysis of the job descriptions showed signicant similarities, and activities specic to scheduling personnel were often listed in the businesss formal documentation. However, the data analysis also provided evidence that schedulers spent a considerable amount of time on nonformalized tasks, in addition to formal tasks. One example was maintenance tasks: Schedulers were often observed arranging their work environment or the data they needed for their job. Sometimes it was simply a matter of rearranging system data into a more suitable format; activities included the development of routines by which schedulers searched for and accessed the data they needed for scheduling. Schedulers also had to compensate in some way for problems or failures in the

542 overall business. These were usually within the information technology system or attributable to the businesss organizational structure. Therefore the results provide evidence that the schedulers undertook three types of task behavior: (a) Formal task behavior is what each company has specied that scheduling personnel should carry out (i.e., the task requirements expected of the scheduling job are formalized and made explicit by the companies). (b) Maintenance task behavior is what the schedulers demonstrate to organize the information they need, such as checking their sources of information, updating the information system or systems, and adjusting data into a suitable format for the task requirements. This task behavior is undertaken in order to keep the scheduling environment up to date and organized as the scheduler sees fit. Different types of maintenance tasks depend on the individual scheduler. (c) Compensation task behavior is what the schedulers demonstrate while attempting to compensate for some level of problem or failure in the overall system for instance, in the information system, in the organizational structure of the company, or in the operations of the production unit. No absolute distinction can be made among these task behaviors, which are interdependent and are frequently carried out concurrently by the schedulers. Tasks can be categorized within any one, two, or three of the task behaviors.

Fall 2004 Human Factors Roles. It is evident that the schedulers activities comprise more than scheduling tasks. A task model on its own provides only a limited view of the human contribution to scheduling; in addition, the essential contribution of roles must be incorporated. A standard denition of a role is any relatively standardized social position involving specific rights and obligations that an individual is expected and encouraged to perform (Jary & Jary, 1991). That is, social roles encapsulate all the norms with which a person occupying a particular status is obliged to comply. Roles are the enablers of the tasks and ll the gap between the formal and informal systems that operate within a business. The results from our study of schedulers demonstrate the types of roles expected of the schedulers as well as what roles they actually hold. It is evident that all of the schedulers carry out coordinating, monitoring, communicating, problem-solving, facilitating, and interfacing types of activities. Our results show that there are three main types of roles held by the schedulers and that they, like the task behaviors, overlap to an extent. The interpersonal role captures the behaviors of developing interpersonal networks, informal bargaining, friendship and favor networks, and mediation. The types of interaction that the schedulers have with other personnel are evident from the results (see Figure 2) and are implicit within the recorded observation and decision making data.

Figure 2. Relative frequency statistics showing source of scheduler interactions and types of media used for interaction.

A NEW MODEL OF SCHEDULING IN MANUFACTURING The analysis demonstrated that all of the participant schedulers had developed their own interpersonal networks that were complementary to the formal networks, such as reporting hierarchies, organizational structures, and functional business groupings. These networks were used to nd and gather specic information at certain times within the scheduling function. Stoop and Wiers (1996) found that when problems arise on the shop oor, schedulers have to know where work originated and where it is headed in order to be able to communicate with personnel about quality problems and delivery times. Schedulers need extensive knowledge and information about specific jobs and shop oor status. Interpersonal activity has been recorded in other scheduling eld studies (Webster, 2001), and interpersonal networks explain how schedulers gather enriched data (McKay et al., 1995a). These networks are normally informal structures, so the people interacting with the scheduler also provide informal, contextual data. The schedulers also utilized informal bargaining, friendship, and favor networks to support their scheduling tasks, the utility of these networks being dependent on the status of the schedulers in their departments. For all the schedulers, status was determined by four key factors: (a) respect other peoples respect for the scheduler and for his or her ability to do the job well; (b) trust peoples trust in the schedulers judgment; (c) friendship whether or not the other people knew the scheduler and liked him or her; and (d) favors the reciprocation of favors, used as a bargaining tool. Our ndings support previous research that also noted that scheduling involves social interaction and issues such as trust and respect (McKay, Buzacott, Charness, & Safayeni, 1992). Other interesting enabling activities were undertaken by the schedulers. All the schedulers worked at a level where the plan is translated into production that is, the interface where the virtual plan becomes a production reality. Therefore the schedulers had to mediate between the various owners of the schedule. The planners owned the virtual plan in relation to the planning task requirement of achieving customer demands. The production personnel owned the physical production in relation to

543

the production task requirement of achieving maximum resource utilization. The schedulers mediated in terms of the schedule to efciently translate the demands of the plan to meet the demands of the shop oor. The informational role captures the behaviors of acting as an information hub, filtering information to the shop oor, and ensuring that information is accessible and visible. It builds upon the infrastructure created by the schedulers interpersonal role; networking and mediation behaviors enable the transfer of information that the schedulers need to carry out their tasks. The communication and transfer of information are evident across the results for all schedulers (see Table 2 and Figures 1 and 2). Schedulers use, update, and pass on all types of information and, so, are information investigators, information users, and information providers. When these behaviors are combined, the schedulers role as an information hub becomes evident. From this viewpoint, a scheduler is a hub within a department, the point of contact for both internal and external inquiries. The scheduler has information for other personnel from other hubs inside and outside the department, and information is distributed across the schedulers department and other departments via this informational network. There is a two-way ow: Other personnel have relevant and current information to pass on, and the scheduler constantly compiles this to ensure that updated information is being distributed within and beyond the department. Frequency counts of the interactions recorded on the observation sheets have been used to show the interactions and the media by which information was transmitted (see Figure 2 for an example). This role behavior supports the ndings by Wiers (1997) and McKay and Wiers (1999) that much of the schedulers task involves information collection and validation because the scheduler needs available and accurate data to make scheduling decisions. Schedulers also lter the information they use. The filtering function describes a range of behaviors, including the selection, smoothing out, verication, interpolation, and transformation of data into formats appropriate to the current scheduling situation. For example, the schedulers were observed adjusting the gures presented

544 in the approximate plan to make them t that is, the computer system aggregated the data, but a person produced the final best fit of the data by verifying or adjusting the figures with respect to the real-world constraints and dynamics of the production situation. Similarly, many pieces of information were not included in the systems data, so it could not generate a feasible, doable plan. The role of the human filter in this context was to fill in information to make a feasible plan possible. This ltering behavior ensures that a real-world production plan is produced in which the scheduler minimizes the possibility of problems reaching the shop oor. The underlying assumption of most computerbased production control systems is that the shop floor situation is stable, all materials are available, the operators are capable of carrying out their jobs, and there are no machine breakdowns. However, when these conditions (or countless others) are not met, schedulers lter the planned information in the form of the schedule through their store of current knowledge. This knowledge includes detailed and up-to-date information regarding the current production situation. If there is not an appropriate match between what is planned and what is occurring on the shop oor, the scheduler attempts to manipulate either the schedule or the production environment in order to implement a solution. Therefore, it is evident that there are two methods of data flow in most manufacturing scenarios: the information system data update and the transfer of knowledge and information by human schedulers to the next level of the process. Within this context, it is obvious that schedulers must have access to current, relevant information and that much of the information they use has to be visible. If information was not directly accessible or visible, then the schedulers worked around the problem. Some schedulers worked in locations that made it difcult for them to access or see information on the shop oor, but their networks and information hub activities enabled them to gain access to or visibility of the information they needed. Visibility of the physical shop oor production process has been noted as important in operations management (Fogarty, Blackstone,

Fall 2004 Human Factors & Hoffman, 1991; Slack, Chambers, Harland, Harrison, & Johnston, 1998), but visibility of information is important to schedulers because they need to be aware of higher-level planning decisions in order to effectively plan ahead and translate medium-range planning decisions into short-range scheduling-related decisions. In line with our results, Hardstone (1991) also noted that information accessibility and visibility are salient features of scheduling. Schedulers need to be able to see and access relevant and critical information to be able to carry out the scheduling task and to make timely scheduling decisions (Nakamura & Salvendy, 1994). Information access was central to the work of Vera, Lewis, and Lerch (1993), who studied the decision making behavior of supervisors in a mail-sorting facility. They found that the supervisors relied on environmental cues to make decisions and that relevant information was gathered by visual scans of the facility. The decisional role captures three types of decision behaviors: problem prediction and problem solving, interruption handling, and resource allocation. It complements the interpersonal and informational roles that provide the networks to access the data and information needed by the schedulers to make their decisions. The observed schedulers all predicted and solved problems; they had to be aware of what potential problems could arise and decide how to address them to minimize disruption to production. This supports McKay and Wierss (1999) ndings that from an operational perspective, scheduling is a problem solving and not a sequencing function. A scheduler must try to anticipate or preempt potential problems and therefore take proactive rather than reactive decision roles. It is apparent that this behavior is linked to a schedulers awareness of the production situation and to his or her understanding of how problems can impact the current schedule. An interruption, in the context of this role, distracts schedulers from their current activity and causes them to divert attention to another activity. Our nding that schedulers act as interruption handlers reects Sanderson and Morays (1990) view that human schedulers exploit opportunities for action interruptions are often opportunities for the schedulers to act and avoid a problem. Rather than being irritations that

A NEW MODEL OF SCHEDULING IN MANUFACTURING distract the schedulers from their work, interruptions are an integral part of their work. A nal decisional role is resource allocation. As noted earlier, from an operations research perspective, the aim of any scheduling activity is the goal-directed allocation of resources over time to perform a collection of tasks. The schedulers all allocated resources in order to meet their scheduling objectives. However, the results from this study clearly show that resource allocation was not the participant schedulers central activity and that their decisions extend beyond pure resource allocations. The schedulers activities are all carried out within a complex and dynamic task environment, and it is this infrastructure that makes the scheduling complex. Without reference to the scheduling environment, resource allocation activities mean little, which is why the concentration on sequencing problems in scheduling theory will always be limited in providing a realistic model of the scheduling problem. The finding that schedulers do not merely solve combinatorial sequencing problems in order to schedule has been noted previously (Stoop & Wiers, 1996; Webster, 2001). The results here go further: They show that resource allocation is not the central function of scheduling in practice. The overall responsibility of the scheduler in terms of scheduling is explained later. Monitoring. In comparison with previous work in process control environments, (Beishon, 1974; Crossman, 1974), within discrete production environments the concept of monitoring has not been widely studied. It is explained here as an integrating activity that allows schedulers to carry out their job, gather accurate information, and make appropriate decisions that is, monitoring can be considered as the scheduling activity that holds everything together. A distinction must be made between monitoring and checking activities, which are interrelated yet different behaviors. Checking is active; schedulers check situations by seeking out information. Monitoring relates to the schedulers being aware of what is going on and what is about to occur in the environment that is, situated data. The situatedness of data describes how particular data are connected to, and affected by, their environment (Behrmann, 1993). All the schedulers used three sources of situated data:

545

1. Mechanistic situated data are hard manufacturing data, such as the availability of stock, the accuracy of system stock figures, and the lead time of special materials. 2. Context situated data are soft manufacturing data, such as knowing the status of the current schedule and the orders about to be scheduled, or knowing if a machine is unreliable and therefore routing work around it whenever possible. These are the background data that surround the schedulers. 3. Exceptional situated data are the data linked to problems, the nonroutine variances. If the schedulers, during their routine monitoring behavior or because of an interruption, are faced with data they have not come across before, a different approach is needed. If possible, the schedulers try to relate the exceptional data to a similar situation, to pattern match the exceptional exception. If this does not work, they try numerous tactics until they reach a satisfactory solution. A schedulers routine monitoring behavior was often interrupted, as previously described. A denition of an interruption in this research is not the same as a disturbance in operations research (Stoop & Wiers, 1996). A disturbance disrupts the plan or schedule; an interruption disrupts the scheduler or warns or informs of a potential problem. It was evident that for all the schedulers the majority of interruptions were positive with regard to work goals; the analysis of interaction data provided evidence that most positive interruptions either warned or made the scheduler aware of a potential or current problem. When a problem did occur, the scheduler needed to deal with the problem in order to return to routine monitoring. The problem was dealt with through their task and role activities, as described previously. Once a problem had been dealt with, the schedulers ensured that the problem had been completely resolved by conrming their actions with other personnel before returning to routine monitoring activity. This feedback behavior is apparent in the decision diagram depicted in Figure 1. External Factors That Inuence Schedulers The results discussed so far describe the activities of the schedulers studied. However,

546 external factors that place the schedulers activities in the context of the wider business environment were also apparent from the data analysis. The explicit acknowledgement of the inuence of contextual factors is missing from many previous operations research models of scheduling (MacCarthy & Liu, 1993). These factors operationalize the results and bridge the gap between scheduling theory and practice. The five interdependent factors presented here were not derived directly from the pattern coding methods. They emerged from comments made by the schedulers and are implicit within the data collected throughout the study. Within a hypothetico-deductive investigation of human scheduling, the factors would represent the variables to be manipulated. In the context of this descriptive, qualitative investigation, they are presented as ve interdependent factors that describe the contextual framework within which the schedulers worked:
1. manufacturing process: the physical elements of schedule implementation (i.e., the production process, raw materials, machines, and tools); 2. organizational structure: the formal and informal elements of a business within which a scheduler works (e.g., the organizational layout and culture and the interrelationships of business functions); 3. planning and scheduling information systems: the computer and management information systems used to generate and implement the business plans and schedules; 4. people: within a manufacturing business the scheduler interacts with a range of personnel to generate and implement the schedule; and 5. performance measures: the types and levels of performance that the scheduler is expected to achieve and the targets set to reect these. In all cases, performance measures directly inuenced the schedulers behavior.

Fall 2004 Human Factors are analyzed together, it is evident that the schedulers presented three distinct types of activities in their jobs: tasks, roles, and monitoring. The important point in terms of understanding the results is that schedulers do not carry out all tasks in the same way all of the time and that their roles also change over time because of the external factors that inuence the scheduling function, including the inherent instability and uncertainty of the business environment. The three activities and the interrelationships among them provide the structure of the model of scheduling. To place the model into the context of different manufacturing scenarios, the ve external influencing factors are added. The framework produced when all of the components are combined is a new model of scheduling in manufacturing. This is presented in Figure 3. An important observation within all of the environments studied is that scheduling, in practice, centers on facilitation and implementation of the schedule. In reality, this means that scheduling rst ensures that the conditions that enable a schedule to be carried out are present (i.e., facilitating) and then accomplishes the schedule (i.e., implementing). When a schedule cannot be implemented, scheduling is the function of facilitating acceptable alternatives and implementing a strategy to complete the planned work. This emphasis on the facilitation and implementation aspects of scheduling also provides an explanation of the gap between scheduling theory and practice. Scheduling theory is based on the classical emphasis on allocation and sequencing decisions and on algorithmic solutions. Scheduling practice places an emphasis, as demonstrated by this research, on social and organizational processes. POTENTIAL APPLICATIONS OF A NEW MODEL OF SCHEDULING Many previous scheduling models, as well as planning and scheduling information systems, are based on the assumption that all schedulers need is an expert or decision support system to help them carry out very dened decision making tasks. Our new model of scheduling suggests that organizational support (e.g., training and communications filtering) is a very important consideration. This change in emphasis from

The five factors of influence, and the interdependencies among them, provide further evidence that schedulers work in complex, dynamic, and context-situated scenarios. A NEW MODEL OF SCHEDULING IN MANUFACTURING One major goal of this research has been to build on the industrial case study ndings in order to draw together the data into a structured framework, presented here as a model of scheduling. When all of the data we have described

A NEW MODEL OF SCHEDULING IN MANUFACTURING

547

Figure 3. A new model of scheduling in manufacturing comprising task, role, and monitoring activities.

scheduling seen as purely an individual cognitive function to it being recognized as a cognitive, social, and organizational function grounded in practice has important implications and applications in business and industry. The model also has potential relevance within the elds of ergonomics and operations management. Scheduling may be a particular specic function (e.g., within a manufacturing business), but it is also a robust reference for the study of complex and dynamic behavior in other work environments and for the implementation of more effective information systems, including enterprise resource planning systems (see Davenport, 1998). Job Analysis There are many potential uses of the model in relation to job analysis. For example, in all of the businesses visited, the management appeared to focus only on the schedulers responsibility to achieve overall task goals and subgoals and had limited understanding of the importance of the schedulers authority, which

enables schedulers to actually implement those responsibilities. As such, authority is a key component that could be built into the design of scheduling jobs. Currently, this is not a common practice within manufacturing businesses. Often, there was also the inaccurate perception that schedulers could be transferred between different departments with limited training. The model suggests that this assumption is likely to be incorrect, as the schedulers performance criteria were context dependent. That is, the scheduling job needs to include the requirement that schedulers have a full and detailed understanding of each particular production environment. In these types of scenarios, a structured job analysis using the model would be a useful asset to a manufacturing business. Performance Metrics Design Effective organizational design of the scheduling function requires an effective performance measurement system. This model provides a framework within which to explicitly determine

548 how to assess scheduling and scheduler performance. In relation to the model, there are two approaches to scheduling performance measures. The case studies demonstrated that formal measures were often found to take the form of production control system-generated measures, originally conceived to measure the operational characteristics of the department. However, the businesses, and the schedulers themselves, often adapted operational performance measures in order to assess scheduler performance. Thus a good scheduler was measured by his or her ability to achieve the numerous, often conicting, criteria by which a schedule is measured. In comparison, the case studies demonstrated that informal measures took the form of the contextual expectations generated by other business personnel. Such performance measures represented the way that schedulers were expected, for example, to be good communicators (interpersonal role), to share accurate and up-to-date information (informational role), and to problem solve and have a proactive view of requirements (decisional role). The model does not propose denitive performance measures of scheduling; instead, it is proposed as a tool to investigate what schedulers actually do in order to dene more appropriate performance metrics and ways of assessing scheduler performance. Job Support Businesses have historically invested heavily in computer-based support mechanisms for scheduling task and monitoring activities, such as expert systems, decision support systems, and advanced planning systems. In relation to supporting scheduler roles, there are fewer support mechanisms in use or even available. Generally, computer-based support focuses on only the decision making component of scheduling, typically resource allocation decisions. The model has shown that this is not the main scheduling function activity. Organization-based support appears to be less common and is often directly linked to business improvement programs, such as business process reengineering exercises. However, organizational support appears to be the more appropriate support mechanism for scheduler roles. This type of support might include regular, proactive meetings related to the state of

Fall 2004 Human Factors the schedule and the shop oor or the clarication by a business of a schedulers function within an organization. Organizational support also encompasses the concept of effective scheduler selection and training approaches. The model provides a framework for understanding the components of the scheduling function in a manufacturing environment. When these components are made transparent to a business, appropriate schedulers can be selected to t the requirements and expectations of a business. Schedulers can then be trained how to work effectively within a particular production control environment. CONCLUSION A new model of scheduling in production environments has been presented here. It is based on extensive empirical study of schedulers in the eld using observation-based methods and decision probes. The model extends considerably the understanding of the nature of scheduling in practice in relation to tasks, roles, and monitoring. It has not been developed as a prescriptive model or a specific scheduling audit tool. The model may help one to understand and evaluate performance measures or to explain and evaluate available computer-based support. It provides a guide to what businesses need to consider and address when designing and assessing planning and scheduling systems, and it helps to clarify the contribution of the human scheduler. Further research is needed to establish how the model may be utilized in the context of both existing and new production planning and control and enterprise information systems. The goal is for practitioners and researchers to use the model of scheduling in order to signicantly enhance how scheduling is viewed and understood in research and in industrial practice. ACKNOWLEDGMENTS This paper stems from a research project at the University of Nottingham titled Effective decision support in production planning and scheduling. The authors would like to express thanks to the Engineering and Physical Sciences Research Council (EPSRC GR/L/31364) for

A NEW MODEL OF SCHEDULING IN MANUFACTURING their support of the research program, to the collaborating companies and the participants, and to the reviewers of the rst version of this paper. REFERENCES
Baker, K. R. (1974). Introduction to scheduling and sequencing. New York: Wiley. Behrmann, L. P. (1993). Representation in situated agents. In Proceedings of the 15th Annual Conference of the Cognitive Science Society (p. 7378). Hillsdale, NJ: Erlbaum. Beishon, R. J. (1974). An analysis and simulation of an operators behavior in controlling continuous baking ovens. In E. Edwards & F. P. Lees (Eds.), The human operator in process control (pp. 7990). London: Taylor & Francis. Bernard, H. R. (1995). Research methods in anthropology: Qualitative and quantitative approaches (2nd ed.). Walnut Creek, CA: AltaMira. Crawford, S., MacCarthy, B. L., Wilson, J. R., & Vernon, C. (1999). Investigating the work of industrial schedulers through field study. Cognition, Technology, and Work, 1, 6377. Crawford, S., & Wiers V. C. S. (2001). From anecdotes to theory: Reviewing the knowledge of the human factors in planning and scheduling. In B. L. MacCarthy & J. R. Wilson (Eds.), Human performance in planning and scheduling (pp. 1544). London: Taylor & Francis. Crossman, E. R. F. W. (1974). Automation and skill. In E. Edwards & F. P. Lees (Eds.), The human operator in process control (pp. 124). London: Taylor & Francis. Davenport, T. (1998). Putting the enterprise into the enterprise system. Harvard Business Review, 76(4), 121133. Dessouky, M. I., Moray, N., & Kijowski, B. (1995). Taxonomy of scheduling systems as a basis for the study of strategic behavior. Human Factors, 37, 443472. Einhorn, H. J., & Hogarth, R. M. (1981). Behavioral decision theory: Processes of judgment and choice. Annual Review of Psychology, 32, 5388. Fogarty, D. W., Blackstone, J. H., & Hoffmann, T. R. (1991). Production and inventory management. Cincinnati, OH: SouthWestern. Hardstone, G. (1991). A simple planning board and a jolly good memory: Time, work organization and production management technology (Student Paper No. 4). Edinburgh, UK: University of Edinburgh, Edinburgh Programme on Information and Communication Technologies. Higgins, P. G. (1999). Job shop scheduling: Hybrid intelligent human-computer paradigm. Unpublished doctoral dissertation, University of Melbourne, Australia. Higgins, P. G. (2001). Architecture and interface aspects of scheduling decision support. In B. L. MacCarthy & J. R. Wilson (Eds.), Human performance in planning and scheduling (pp. 245281). London: Taylor & Francis. Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: MIT Press. Jary, D., & Jary, J. (1991). Dictionary of sociology (3rd ed.). Glasgow, UK: Unwin Hyman/Harper Collins. Keeney, R. L., & Raiffa, H. (1976). Decisions with multiple objectives: Preferences and value trade-offs. New York: Wiley. Kirwan, B., & Ainsworth, L. K. (1992). A guide to task analysis. London: Taylor & Francis. Klein, G. A., Orasanu, J., Calderwood, R., & Zsambok, C. E. (1993). Decision making in action: Models and methods. Norwood, NJ: Ablex. Lipshitz, R., Klein, G., Orasanu, J., & Salas, E. (2001). Taking stock of naturalistic decision making. Journal of Behavioral Decision Making, 14, 331352. Lipshitz, R., & Strauss, O. (1997). Coping with uncertainty: A naturalistic decision making analysis. Organizational Behavior and Human Decision Processes, 69, 149163. MacCarthy, B. L., & Liu, J. (1993). Addressing the gap in scheduling research: A review of optimization and heuristic methods in production scheduling. International Journal of Production Research 31, 5979.

549

MacCarthy, B. L., & Wilson, J. R. (2001). Human performance in planning and scheduling. London: Taylor & Francis. MacCarthy, B. L., Wilson, J. R., & Crawford, S. (2001). Human performance in industrial scheduling: A framework for understanding. Human Factors and Ergonomics in Manufacturing, 11(4), 122. McKay, K. N., Buzacott, J. A., Charness, N., & Safayeni, F. R. (1992). The schedulers predictive expertise: An interdiciplinary perspective. In G. I. Doukidis & R. J. Paul (Eds.), Artificial intelligence in operational research (pp. 139150). Basingstoke, UK: Macmillan. McKay, K. N., Safayeni, F. R., & Buzacott, J. A. (1995a). Common sense realities of planning and scheduling in printed circuit board manufacture. International Journal of Production Research, 33, 15871603. McKay, K. N., Safayeni, F. R., & Buzacott, J. A. (1995b). Schedulers and planners: What and how can we learn from them. In D. E. Brown & W. T. Scherer (Eds.), Intelligent scheduling systems (pp. 4162). London: Kluwer. McKay, K. N., & Wiers V. C. S. (1999). Unifying the theory and practice of production scheduling. Journal of Manufacturing Systems 18, 241255. Miles, M. B. & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook. Thousand Oaks, CA: Sage. Milliken, F. J. (1987). Three types of perceived uncertainty about the environment: State, effect and response uncertainty. Academy of Management Review, 12, 133143. Morton, T. E., & Pentico, D. W. (1993). Heuristic scheduling systems. New York: Wiley. Nakamura, N., & Salvendy, G. (1994). Human planner and scheduler. In G. Salvendy & W. Karwowski (Eds.), Design of work and development of personnel in advanced manufacturing (pp. 331354). New York: Wiley. Orasanu, J., & Connolly, T. (1993). The reinvention of decision making. In G. A. Klein, J. Orasanu, R. Calderwood, & C. E. Zsambok (Eds.), Decision making in action: Models and methods (pp. 320). Norwood, NJ: Ablex. Ovacik, I. A., & Uzsoy, R. (1997). Decomposition methods for complex factory scheduling problems. Boston, MA: Kluwer Academic. Pinedo, M. (1995). Scheduling: Theory, algorithms and systems. Englewood Cliffs, NJ: Prentice-Hall. Prietula, M. J., Carley, K. M., & Gasser, L. (1998). Simulating organizations: Computational models of institutions and groups. Cambridge, MA: MIT Press. Robson, C. (1993). Real world research: A resource for social scientists and practitioner-researchers. Oxford, UK: Blackwell. Sadowski, R. P., & Medeiros, D. J. (1982). Planning and control for manufacturing systems and projects. In G. Salvendy (Ed.), Handbook of industrial engineering (pp. 11.2.111.2.24.). New York: Wiley. Sanderson, P. M. (1991). Towards the model human scheduler. International Journal of Human Factors in Manufacturing, 1, 195219. Sanderson, P. M., & Moray, N. (1990). The human factors of scheduling behavior. In W. Karwowski & M. Rahimi (Eds.), Ergonomics of hybrid automated systems II (pp. 399406). Amsterdam: Elsevier. Slack, N., Chambers, S., Harland, C., Harrison, A., & Johnston, R. (1998). Operations management (2nd ed.). London: Financial Times-Pitman. Slovic, P., Fischoff, B., & Lichenstein, S. (1977). Behavioral decision making. Annual Review of Psychology, 28, 139. Stammers, R. B., & Shepherd, A. (1995). Task analysis. In J. R. Wilson & E. N. Corlett (Eds.), Evaluation of human work (2nd ed., pp. 144168). London: Taylor & Francis. Stoop, P. M., & Wiers, V. C. S. (1996). The complexity of scheduling in practice. International Journal of Operations and Production Management, 16(10), 3753. Strauss, A. L. (1987). Qualitative analysis for social scientists. Cambridge, UK: Cambridge University Press. Strauss, A. L., & Corbin, J. (1994). Grounded theory methodology. In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 273285). Thousand Oaks, CA: Sage. Vera, A. H., Lewis, R. L., & Lerch, F. J. (1993). Situated decision making and recognition based learning: Applying symbolic

550
theories to interactive tasks. In Proceedings of the 15th Annual Conference of the Cognitive Science Society (pp. 8495). Hillsdale, NJ: Erlbaum. Webster, S. (2001). A case study of scheduling practice at a machine tool manufacturer. In B. L. MacCarthy & J. R. Wilson (Eds.), Human performance in planning and scheduling (pp. 6782). London: Taylor & Francis. Wiers, V. C. S. (1996). A quantitative field study of the decision behavior of four shop floor schedulers. Production Planning and Control, 7, 383392. Wiers, V. C. S. (1997). Human computer interaction in production scheduling: Analysis and design of decision support systems in production scheduling tasks. Unpublished doctoral dissertation, University of Eindhoven, Netherlands.

Fall 2004 Human Factors


John R. Wilson is a professor and director of the Institute for Occupational Ergonomics at the University of Nottingham. He received his Ph.D. in work design and ergonomics from the University of Birmingham in 1983. Bart L. MacCarthy is a professor of operations management, head of the Operations Management Division at Nottingham University Business School, and director of the Mass Customization Research Centre. He received his Ph.D. in manufacturing systems from the University of Bradford in 1986. Date received: April 25, 2002 Date accepted: January 12, 2004

Sarah Jackson (nee Crawford) is a research fellow in the Institute of Occupational Ergonomics at the University of Nottingham. She received her Ph.D. in operations management from the University of Nottingham in 2000.

Das könnte Ihnen auch gefallen