Beruflich Dokumente
Kultur Dokumente
a r t i c l e
i n f o
Article history:
Received 11 December 2014
Revised 26 January 2016
Accepted 27 January 2016
Available online 13 February 2016
Keywords:
Fuzzy logic
System dynamics
Balance dynamic scorecard
Managerial dashboard
a b s t r a c t
This paper aims to show how the use of fuzzy systems can enhance the application of system dynamics
(SD), with the construction of virtual worlds, for organizational learning. By doing so, the main contribution is to propose a fuzzy-SD integrated methodology that allows a natural language modeling of decision
policies. At rst it will be revised the recent literature with practical and theoretical fuzzy-SD integration
showing that the motivation and purpose of this article have not been explored yet. Following, it will
be scrutinized the origin of system dynamics to theoretically justify the proposed fuzzy-SD integration
for organizational learning. Finally, it will be presented a hypothetical case study based on the balance
scorecard model to illustrate the methodology in action.
2016 Elsevier Ltd. All rights reserved.
1. Introduction
Almost simultaneously, two revolutionary knowledge elds
were developed: Systems Dynamics, created by Jay Forrester at
Massachusetts Institute of Technology in the 50s (Forrester, 1961),
and Fuzzy Logic created by Lot Zadeh in the 60s (Zadeh,
1975a). Both arose from the perception that the tools of their times
were unable to deal appropriately with real problems (Sterman,
20 0 0; Ross, 2005).
System Dynamics (SD) can be dened as the branch of control
theory that deals with socioeconomic systems and the branch of
science that deals with management issues controllability (Coyle,
1996). Forrester perceived the deciency in our decision-making
process when dealing with the complexity of dynamic systems. To
overcome this deciency, he proposed modeling tools that allowed
discussion, rening and simulation of mental models exposed to
different policies. His greatest contribution was the Stock and Flow
Language that allowed the creation of virtual worlds, controlled experimental laboratories to enhance learning.
Fuzzy Logic (FL) and related Fuzzy Sets Theory have been created with the purpose of adapting the mathematical tools of logic
to types of uncertainty such as vagueness and approximation, characteristics of natural language and of human mental models. By
doing so, FL enables a representation of human knowledge by lin
Corresponding author.
E-mail addresses: daniel_cardoso@poli.ufrj.br (D.C. de Salles),
armandoneto@poli.ufrj.br (A.C. Gonalves Neto), lgmarujo@poli.ufrj.br (L.G. Marujo).
1
Tel.: +55 21 3938 8065.
http://dx.doi.org/10.1016/j.eswa.2016.01.048
0957-4174/ 2016 Elsevier Ltd. All rights reserved.
2. Review of literature
Traditionally, literature uses the FL approach especially for natural language processing and inaccurate knowledge in expert systems, process control and pattern recognition (Karavezyris, Timpe,
& Marzi, 2002). At the other extreme, the literature of dynamic
systems permeates many areas of study, especially socioeconomic
and administrative, basically applicable to any system we wish to
study.
The rst authors to integrate these two approaches were Pankaj
and Sushil (1994), who proposed a method for qualitative analysis
of causal loops using fuzzy linguistic uncertainties to incorporate
the perceptions and beliefs of the modeler. As in most of later integration proposals their motivation was the perception that variables relationships in humans mental models are best expressed
in natural language.
173
174
Control
Indicators
Policys
hypotheses
System
dynamics
Model
Fuzzy policy
system
Dynamics
hypotheses
Control
Decisions
Fig. 1. Scheme of the integration of fuzzy system and SD model.
1. Problem Articulation
(Boundaries Selection)
1.a Theme Selecon
1.b Key Variables
1.c Time Horizon
1.d Dynamic problem denion
2. Dynamic
Hypotheses
2.a Inial hypothesis generaon
2.b Endogenous focus
2.c Mapping the structure
5. Formulation and
Evaluation of Fuzzy
Policies 5.a Set controlled variables
5.b Dene indicators and frequency
5.c Set Mamdani system
3. Formulation
4. Testing
4.a Comparison with reference modes
4.b Robustness under extreme
condions
4.c Sensivity
4. Proposed methodology
5. The case study: a hypothetical organization
In the proposed methodology, the system to be governed (modeled on SD stock and ow language) is separated from the human
decisions system (policies), which can now be modeled in a natural language using the fuzzy Mamdani system (Mamdani & Assilian, 1975) and formally declared, facilitating the application of the
scientic method and disciplining policies testing. Fig. 1 illustrates
this interaction.
This methodology can be included in a more general framework
called Adaptive Control Methodology (ACM) developed by Brans,
Kunsch, and Mareschal (2002); Brans, Macharis, Kunsch, Cheva-
175
Investment
+
Financial
Profit
Revenue
+ Price
Demand
Customers
-
Delivery Time
Internal
Processes
+ Propensity to
consumption
Production
Capacity
Costs
+
+
+
Productivity
Staff
Learning
And Growth
+
Processes
Improvement
Innovations
is assumed that the initial problem arose from previous innovation investments that made
the product attractive for
the market (considering its low price) and made the demand
to exceed the production capacity.
(c) Mapping system structure:
Fig. 3 shows the causal loop diagram
balance scorecard structure.
aligned with
the
176
177
Table 1
Equations explanation for each variable in the stock and ow model.
FORMULAE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
OBSERVATIONS
Dt = 0.5
Innovation_Investment_percentual = 50 0 0/150 0 0
Innovation(t) = Innovation (tdt) + (Innovation_Investment) dt
INIT Innovation = 50 0 0
Innovation_Investment = (Investment / Innovation)
Innovation_Investment_percentual
Processes_Improvement (t) = Processes_Improvement
(tdt) + (Processes_Improvement_Investment) dt
INIT Processes_Improvement = 10 0 0 0
Processes_Improvement_Investment = (Investiment/Processes_Improvement)
(1Innovation_Investment_percentual)
Price(t) = Price (tdt) + (Delta_Price) dt
INIT Price = 50
Delta_Price = 0
Production(t) = Production (tdt) + (Orders - Expedition) dt
INIT Production = 0
Orders = 10 0 0 0 Propensity_to_ consumption / 100
Expedition = Productivity Staff
Staff(t) = Staff(tdt) + (Delta_Staff) dt
INIT Staff = 20
Delta_Staff = 0
Labor_Cost = Staff 2500
Material_Cost = 5
Total_Cost = (Expedition Material_Cost) + Labor_Cost
Investment = 0.5 Prot
Prot = RevenuesTotal_Cost
Productivity = (450 Processes_Improvement)/10 0 0 0
25
26
27
Delivery_Time = CTMEAN(Expedition) 30
Capacity_Utilization = (Orders/Capacity) 100
28
29
Revenue_Variation = DERIVN(Vendas,1)
Delta_Revenue = IF Revenue = 0 THEN 0 ELSE (Revenue_Variation/Revenue)
100
Revenue = Expedition Price
NPV = NPV(Prot,0.005)
30
31
(17/
178
assuming that at some point there would be an ideal balance between price level and team size that would not allow
more signicant sales growth.
(c) Dene fuzzy controller or fuzzy policy system:
The fuzzy system best suited to implement rules in natural language is the Mamdani system (Mamdani & Assilian 1975). To
build such system it must be considered the following items:
179
make sense or viability in real life, because you can lose all practical value of learning.
We dened, for the price control, a $10 range of variation. The
changes of staff, for reasons merely subjective, was set to layoff not
exceeding ve employees and hire not above ten employees (in an
interval of 30 days).
As can be noted in Fig. 9 the Capacity Utilization was considered medium close to 70% (steeper for larger values) and this will
be the control target.
Fig. 10 highlights the range mentioned and price maintenance
between $2. In Fig. 11, it can be seen that the Delta Revenue was
dened very sensitive to the variations, considering null between
180
case, growing at a faster pace than in the baseline behavior, although the graph scale does not allow the comparison
to Fig. 5.
At the nancial dimension, it should be noted that the total
cost now behaves desynchronized with prot and sales, because its
xed portion changes with the workforces hiring and ring. Also,
note that the prot presented an increasing pattern with the
increase in the propensity to consume and the price level. For
comparison, the NPV of this model was $1.132 107 , a result considerably higher than the model without control (shown in the
sensitivity analysis, Fig. 8).
Interestingly, after achieving half of the simulation, the control
was stabilized keeping the price around $66, and a workforce of 27
employees.
This hypothetical case can give some managerial insights on
companies releasing new innovative products. As usual the product is initially offered at a relative low price (with low prot margin) to foment demand, however as the demand grows at fast rate
the enterprise should increase their capacity (in this case related
uniquely to staff size) to avoid deterioration in delivery service.
Parallel to the capacity increase, this study suggest that the price
could be raised gradually (as the innovation and good delivery service guarantee an eager demand), until a stabilization level where
prot margins are optimized without reducing the demand for the
181
Capacity utilization IS
1
2
3
4
5
6
7
8
9
Low
Low
Low
Medium
Medium
Medium
High
High
High
AND
Delta revenueIS
Positive
Null
Negative
Positive
Null
Negative
Positive
Null
Negative
THEN
Delta price IS
Delta staffIS
Keep
Reduce
Reduce
Keep
Keep
Reduce
Raise
Raise
Raise
Reduce
Reduce
Reduce
Reduce
Keep
Keep
Raise
Raise
High Raise
182
References
Arasteh, A., & Aliahmadi, A. (2014). A proposed real options method for assessing
investments. The International Journal of Advanced Manufacturing Technology, 70,
13771393.
Baradaran, V., & Keshavarz, M. (2015). An integrated approach of system dynamics simulation and fuzzy inference system for retailers credit scoring. Economic
Research-Ekonomska Istraivanja, 28, 959980.
Brans, J. P., Macharis, C., Kunsch, P. L., Chevalier, A., & Schwaninger, M. (1998). Combining multi-criteria decision aid and system dynamics for the control of socio-economic processes: an iterative real-time procedure. European Journal of
Operational Research, 109, 428441.
Brans, J. P., Kunsch, P. L., & Mareschal, B. (2002). Management of the future: a system dynamics and MCDA approach. In D. Bouyssou, E. Jacquet-Lagrze, P. Perny,
R. Slowinski, D. Vanderpooten, & Ph. Vincke (Eds.), Aiding Decisions with Multiple Criteria: Essays in Honour of Bernard Roy (pp. 483502). Kluwer Academic
Publishers, Dordrecht, the Netherlands.
Campuzano, F., Mula, J., & Peidro, D. (2010). Fuzzy estimations and system dynamics
for improving supply chains. Fuzzy Sets and Systems, 161, 15301542.
Chang, Y., & Ko, T. (2014). An interactive dynamic multi-objective programming
model to support better land use planning. Land Use Policy, 36, 1322.
Coyle, R. G. (1996). System dynamics modelling: a practical Approch. CRC Press.
Forrester, J. W. (1961). Industrial dynamics. Cambridge, MA: MIT Press.
Forrester, J. W. (1987). 14 obvious Truths. System Dynamics Review, 3, 156159.
Gary, M. S., Kunc, M., Morecroft, J. D. W., & Rockart, S. F. (2008). System dynamics
and strategy. System Dynamics Review, 24, 407429.
Ghemawat, P., & Cassiman, B. (2007). Introduction to the special issue on strategic
dynamics. Management Science, 53, 529536.
Inc, TheMathWorks (2013). MATLAB. Natick, Massachusetts, U.S.A.
ISEE systems. (2009). STELLA. Version 9.1.
Karavezyris, V., Timpe, K. ,, & Marzi, R. (2002). Application of system dynamics and
fuzzy logic to forecasting of municipal solid waste. Mathematics and Computers
in Simulation, 60, 149158.
Khanzadi, M., Nasirzadeh, F. ,, & Alipour, M. (2012). Integrating system dynamics and
fuzzy logic modeling to determine concession period in bot projects. Automation
in Construction, 22, 368376.
Kleijnen, J. P. C. (1995). Sensitivity analysis and optimization of system dynamics
models: regression analysis and statistical design of experiments. System Dynamics Review, 11, 275288.
Kunsch, P. L., Springael, J., & Brans, J. P. (2001). An adaptive control methodology
based on system dynamics and MCDA case study: the CO2 energy tax in the
residential sector. In A. Colorni, M. Paruccini, & B. Roy (Eds.), Aide Multi Critre
la Dcision/Multiple Criteria Decision Aiding, Report of the Commission of the
European Communities, Joint Research Centre JRC-ISPRA, EUR - report 19808 EN
(pp. 145162).
Kunsch, P. ,, & Springael, J. (2008). Simulation with system dynamics and fuzzy reasoning of a tax policy to reduce CO2 emissions in the residential sector. European Journal of Operational Research, 185, 12851299.
Mamdani, E. H., & Assilian, S. (1975). An Experiment in linguistic synthesis with a
fuzzy logic controller. International Journal of Man-Machine Studies, 7, 113.
Morecroft, J. D. W. (1988). System dynamics and microworlds for policymakers. European Journal of Operational Research, 35, 301320.
Nasirzadeh, F., Khanzadi, M. ,, & Razaie, M. (2014). Dynamic modeling of the quantitative risk allocation in construction projects. International Journal of Project
Management, 32, 442451.
Oliva, R. (2003). Model calibration as a testing strategy for system dynamics models.
European Journal of Operational Research, 151, 552568.
Orji, I. J., & Wei, S. (2015a). An innovative integration of fuzzy-logic and systems
dynamics in sustainable supplier selection: a case on manufacturing industry.
Computers & Industrial Engineering, 88, 112.
Orji, I. J., & Wei, S. (2015b). Dynamic modeling of sustainable operation in green
manufacturing environment. Journal of Manufacturing Technology Management,
26, 12011217.
Pankaj, S., & Sushil, K. (1994). A fuzzy set theoretic approach to qualitative analysis
of causal loops system dynamics. European Journal of Operational Research, 78,
380393.
Polat, S., & Bozdag , C. F. (2002). Comparison of fuzzy and crisp systems via system
dynamics simulation. European Journal of Operational Research, 138, 178190.
Porter, M. E. (1991). Towards a dynamic theory of strategy. Strategic Management
Journal, 12, 95117.
Richardson, G. P. (2011). Reections on the Foundations of System Dynamics. System
Dynamics Review, 27, 219243.
Ross, T. J. (2005). Fuzzy logic with engineering applications. Hoboken, NJ, USA: John
Wiley & Sons.
Sabounchi, N. S., Triantis, K. P., Sarangi, S., & Liu, S. (2014). Dynamic simulation modeling and policy analysis of an area-based congestion pricing scheme
for a transportation socioeconomic system. Transportation Research Part A, 59,
357383.
Saleh, M., Oliva, R., Kampmann, C. E., & Davidsen, P. I. (2010). A comprehensive analytical approach for policy analysis of system dynamics models. European Journal
of Operational Research, 203, 673683.
Song, J., Song, D., & Zhang, D. (2015). Modeling the concession period and subsidy
for bot waste-to-energy incineration projects. Journal of Construction Engineering
and Management, 141, 04015033.
Sterman, J. D. (20 0 0). Business Dynamics: Systems Thinking and Modeling for a Complex World. Irwin/McGraw-Hill.
183
Xu, J., Deng, Y., & Yao, L. (2014). Sustainable development-oriented industrial restructuring modeling and analysis: a case study in Leshan. Clean Technologies
and Environmental Policy, 16, 267279.
Zadeh, L. A. (1975a). The concept of a linguistic variable and its application to approximate reasoning-I. Journal of Information Science, 8, 199245.
Zadeh, L. A. (1975b). The concept of a linguistic variable and its application to approximate reasoning-II. Information Sciences, 8, 301357.