Sie sind auf Seite 1von 4

An online performance evaluation of business partners system: OLAP based approach

Department of Industrial & Systems Engineering The Hong Kong Polytechnic University Hung Hom HONG KONG mfhenry@,dinet.polyu. edu.hk
Abstract -The core of manufacturing activities has shifted from the physical production of goods to the systematic processing of knowledge to create value for the customers from shortened product development cycle, integrated enterprise modelling, improved outbound logistics and fast response to market needs using innovative information tools and digital technology in the network environment. Manufacturing-related business services that capitalize on information technology are expanding fast. The competitiveness of tomorrow's manufacturing enterprise lies in the development and adoption of a wide range of digital technologies and tools and the creation of virtual and fast-response enterprise to meet the new industrial order of the next century. The formation of digital manufacturing networks is taking up momentum to meet this challenge. In this new business model of virtual enterprise, the partnership synergy is an important issue to be addressed. This paper attempts to provide a model to analyze the issue of partner synergy in a digital manufacturing network. In particular, a model to evaluate the performance of partners in "real time" based on the OnLine Analytical Processing (OLAP) technology has been proposed. In this paper, a case example has also been covered to evaluate the feasibility of this approach.

H.C.W. LAU

Department of Industrial & Systems Engineering The Hong Kong Polytechnic University Hung Hom. HONG KONG 039001OOr@,polyu. edu.hk

G.T.S. HO

implementing Online Analytical Processing (OLAP) on 'Factory On Demand', the data needed for automatic selection of partners with minimum intervention can be achieved.. II. WHAT IS OLAP?
Data mining is a technology that provides sophisticated analysis for applications such as data warehouse [4,5,6,7]. Data mining tools require different data formats in relational and multi-dimensional database systems. The shared data access interface of data mining tools will enable easier exchange of information among different sources. A typical example of data mining tool is On-line Analytical Processing (OLAP) which provides a service for accessing, viewing and analyzing on large volumes of data with high flexibility and performance. The essential characteristic of OLAP is that it enables a numerical and statistical analysis of data which is organized in multi-dimension. OLAP provides quick responses and accurate result to the user, thereby improve the application security and user interface [8].

I. INTRODUCTION
In the past decade, the change in the global economy has entirely redefined indus-trial competitiveness. The core activity of manufacturing is no longer confined to making things but lies in the systematic processing of knowledge to create value for customers. Manufacturing-related business services that capitalize on Internet and information technologies are expanding fast. In fact, the availability of digital tools and equipment in the marketplace provide a major opportunity for enterprises to build stronger relationships with customers and partners which is regarded as the key to building a successful business. A model of collaborative manufacturing network known as Factory On Demand (FOD) is proposed by the authors [1] to link up the manufacturing resources of small & medium firms to form dynamic virtual enterprise for specific tasks. Various services from industrial design, product modelling and analysis, rapid prototyping and tooling can be outsourced and sought from the FOD network. Among the many technical & organizational problems to be solved, the selection of virtual partners in such a dynamic network is of prime importance. Partner selection is a major process of business operation [2,3]. With a good partner selection, the supplier may provide a good price, quality and delivery status so the efficiency of enterprise business can be improved. This paper presents a method of the partner selection based on price, quality and delivery with the 'Factory On Demand' [1] case. By

III. FEATURES OF MULTI- DIMENSIONAL DATA MODEL


An OLAP data model consists of descriptive data (dimensions) and quantitative value (measures) which generates the OLAP data cube [9] by building up two ele-ments: fact table and dimension (an example to illustrate the data cube is covered in the following section). In the fact table, the required data and user-defined method for analysis is specified clearly. In the dimension, the different dimension levels are clearly defined for using on different views of OLAP data cube. Typical dimension includes location, company and time whereas typical measure includes price, sales and profit. In multi-dimensional data model, data is organized in a hierarchy that represents different levels of the data. This distinct feature can be used to compute a complex query and analyze data on reports, thereby achieving the switch of views of data in different dimensions more easily. IV. DIFFERENCE BETWEEN TRADITIONAL APPROACH AND OLAP In the traditional approach, when a user needs to retrieve information across multi-tables, he/she must clearly define the necessary tables for finding the specific information, e.g. Microsoft Access. When a user needs to know how much sales is taken for a certain year & city and uses Internet as

1139

0-7803-9701-0/06/$20.00 (C2006 IEEE

promotion media, the tables and their relationships must be clearly determined
to retrieve information from the tables

Then, the Structured Query Language (SQL) is constructed as below.

Before the implementation of OLAP, the calculated member that used in partner selection is constructed by the measures. The method of how to find suitable partners is described in Figure 3.

" Select sum(a.store_sales) from sales_fact_1997 a, promotion b, region c, store d where a.store id = d.store_id and c.region_id = d.region_id and b.promotion id = a.promotion_id and c.sales_country USA and b.media_type = Internet"

[ta ~ ~
OellAvefaemg eelSo
+

[:A:::ae

crs o

,;

fas ervelPoresov uaityd e

ve

By using the OLAP approach, the table used for the query and the data used to perform the calculation are defined in the dimension and measures separately. Then, the user builds up a complex calculation on calculated member to meet the specific requirements as shown in Figure 2.

1JbWeo

la'

Figure 3: Partner scoring method


In the above diagram, the scores of delivery level, quality level and price level of each job is defined by the measures. Then, the average score of a job is calculated by average scores of the delivery level, price level and quality level of the previous jobs that had been done by service provider. Finally, the overall average of the service provider is determined by accumulating the different weighting of the latest 4 jobs in maximum. Table 1 shows the weight of average score of the job
Job Available

Weighting Method

1gU1i.,

Because of the calculation and analyzes have been pre-computed in OLAP server previously, only a simple multi-dimension expression (MDX) is necessary to construct for retrieving identical result as shown in the following.

3
4 or above

" Select [Measures].[Store Sales] [Store]. [Sales_country] on rows from ([Promotion].[Media Type]. [Internet], [Region]. [Sales-country]. [USA])"

on

sales

columns,

Average Scores of Job x 100% Average Scores of Jobl x 75% Scores of Job2 x 25% Average Scores of Jobl x 50% Scores of Job2 x 30% + Average Scores of Job3 x 20% Average Scores of Jobl x 40% Scores of Job2 x 30% + Average Scores of Job3 x 20% Scores of Job4 x 10%
-

+ Average + Average + Average + Average

where

Deviation Percentage =Quoted Price Required Price Required Price


Table 1: Job weighting method
5.1 Score for Price

x 100%

In the above expression, it can be shown that MDX is more simple and clear than SQL statement. When the user requirement is changed, only minor part of the OLAP data cube and simple multi-dimension expression are needed to be modified to fulfil the user requirement. In traditional approach, the SQL statement may be rewritten to fulfil a new requirement. So, OLAP approach is more efficient and easier to implement than the traditional method. In general, MDX is suitable for decision support system. The typical example of decision support system as follows.
> IF([Measures].[Unit Sales] 1000, [Measures].[Store_Sales] * 0.9, [Meas-ures].[Store_Sales])".

In the above formula, the "Deviation Percentage" is determined by the customer required price and service provider quoted price. When "Deviation Percentage" gets a negative result, it means that the service providers can provide a lower price level for their job. In Table 2, the score of price level is determined by "Deviation Percent-age" of the price level.
Deviation Percentage Below -25% Between -25% and -15% Between -15%
Level

Score
7 marks 6 marks 5 marks

V. THE IMPLEMENTATION OF MULTI-DIMENSION FEATURES ON PARTNER SELECTION

Extremely Low Very Low Fairly Low

1140

2006 IEEE International Conference

on

Industrial Informatics

and -5% Between -5% and

Average Fairly High Very High

4 marks

Between 1%

and

3 marks
2 marks

1 .75marks Question 3 7marks/4 Question 4 1 .75marks 7marks/4 Table 4: Quality level score mapping table VI. REAL-LIFE IMPLEMENTATION OF OLAP
TFCHNCT.C)GV

Between 15% and

Above 25%

Extremely High

1 marks

5.2

Table 2: Price level score mapping Score for Delivery


Deviation Percentage Level
Extremely Low Very Low Fairly Low
Average

Score
7 marks

Below -25%

Between -25% and -15% Between -15% and -5% Between -5% and 5% Between 5% and 15% Between 15% and 25% Above 25%

6 marks 5 marks 4 marks 3 marks 2 marks

Fairly High
Very High

Figure 8: Visual Basic application using OLAP technology


MDX is not just used retrieve analytical result by multi-dimensional expression (MDX) query. It can be embedded in some Windows based development tools such as Visual Basic, Visual Fox Pro, Power Builder and etc. In the Figure 8, it is a simple Visual Basic program that uses embedded MDX to find a suitable partner based on specific requirements.

Table 3: Delivery level score mapping table


In the above formula, the "Deviation Percentage" is determined by the actual delivery date, the proposed delivery date and the days needed for the job. When "Deviation Percentage" gets a negative result, it means that the service providers deliver the product/service on or before the proposed date. Similarly, a positive result signifies an on time delivery. In Table 3, the score of delivery level is determined by "Deviation Percentage" of the delivery level.

High

Extremely

1 marks

5.3

Score for Quality

In the quality level scores calculation; the questionnaire is used to collect the quality level from the customer through Internet. Therefore, the quality level scores will be determined by sum up each positive answer from the questions below and get the quality level scores from Table 4. Does the service provider have a proper customer support procedures for their product? Does the service provider deliver a product in your expected quality? Does the service provider afford a good communication within the processing procedure? Does the service provider give sufficient information for you to monitor the job?

Because OLAP technology can be used on different development tools, it can be transferred the above simple VB program on the partner selection of FOD. In FOD, after the customer submits the requirement to FOD server, the OLAP technology can be used to find available service providers for the customer based on the core competence specify. In the Figure 9, the system lists out available service providers for the customer to decide which company is more suitable for them.The authors gratefully acknowledge the contributions of Ian X. Austan, Ann Burgmeyer, C.J. Essel, and S. H. Gold.

Question Number
Question 1 Question 2

Calculation Method 7marks/4 7marks/4

Score
1.75marks 1.75marks

,AP technology on partner selection of FOD

2006 IEEE International Conference

on

Industrial Informatics

1141

|_
Sbi'Alnr ProvMde: Gu3ngzhouj Erat.rpre

[4] Berson, A., Data Warehousing, Data Mining and OLAP, ~McGraw-Hall, 1997.
[5] Michael, L. G. and Bel, G. R., Data mining - a powerful information creating tool, OCLC Systems & Services, | 15:2, 1i llVol. 1999, pp8 -90.

AM

3
|

4.0

[6] Robert S. C., Joseph A. V. and David B., Microsoft Data > ,''Warehousing, John Wiley & Sons, 1999.

......

qB 1!191 | 21

s ii|,l

ffi

gl'' "*|Sl' l[7]Timothy, P., f|t(


.. 2# I

Jim, P. and Russell D., Microsoft OLAP Unleashed SAMS Publishing 1999.

_____________X_____________________ , Figure 10: History Jobs Information

[8] Margarent, A. H. and Rod, H., Facilitating corporate knowledge: building the data warehouse, Information &

Computer Security, Vol.5:5, 1997, ppl70-174.

And then, the detailed history job information may be displayed by clicking the particular company name as shown in figure 10. Otherwise, it can be submitted E-mail to these service providers automatically by clicking the "E-mail checked items" button.
VII. CONCLUSION AND FUTURE WORKS

[9] Erik, T., George, S. and Dick, C., Microsoft OLAP Solutions, John Wiley & Sons, 1999.

Sum up the above opinions, OLAP approach may be more suitable to be used on partner selection of FOD than traditional approach such as MS Access because OLAP approach can provide multi-dimension views of data and can be used to produce a intermediate result to perform calculation. When the OLAP approach is ap-plied on FOD, the table with analyzed data will help the customer on decision making process. FOD will use the ActiveX control called "Agent OCX" to assist OLAP technology to perform predication and forecast the future behaviour of service providers and monitoring the work in progress. Finally, FOD will use XML to assist the original HTML and ASP on the Web development and uses its features on electronic data interchange.
VIII. REFERENCES

[1] Lee, W.B. and Lau, H.C.W., Factory on Demand: the shaping of an agile production network, International Journal of Agile Management Systems, Vol.1:2, 1999, pp83-87.

[2] William, J. B., Ron, M. and Robert, R., Choosing successful technology development partners: a best-practice model, Vol. 15:1/2, 1998, ppl24-138.

[3] Jaideep, M., Mohamed, Y., Yunus, K. and Elizabeth, F., Supplier selection in developing countries: a model development, Integrated Manufacturing Systems, Vol.10:3, 1999, ppl54-162.

1142

2006 IEEE International Conference on Industrial Informatics

Das könnte Ihnen auch gefallen