Optimal fuzzy multi-criteria expansion of competence sets using multi-objectives evolutionary algorithms
Introduction
Making good decisions involve the successive accumulation of the particular skills, ideas, information and knowledge. In order to efficiently and effectively acquire these abilities, competence set analysis was proposed (Yu, 1990, Yu and Zhang, 1990). Using the methods of competence set analysis such as the minimum spanning tree (Yu & Zhang, 1992) or the mathematical programming (Shi & Yu, 1996), we can obtain the optimal path (e.g. the minimum cost or time) to acquire the required competence.
In conventional competence set analysis, one criterion such as cost or benefit function is used to select the optimal expansion process. However, in practice we usually determine the optimal expansion process according to multi-criteria (e.g. cost, time, efficient, benefit, and so on) simultaneously. Additionally, in order to reflect the ambiguity and uncertainty in practice, we should incorporate the concept of fuzzy sets into competence set analysis.
In order to deal with multi-criteria problems, many methods such as goal programming (Charnes and Cooper, 1957, Ijiri, 1965), min–max optimization (Osyczka, 1978, Rao, 1986) and the ε-constraint method (Osyczka, 1984, Hwang et al., 1980) have been proposed. Recently, MOEA has been widely used in various multi-objective problems such as scheduling (Murata, Ishibuchi, & Tanaka, 1996), engineering (Fonseca & Fleming, 1998) and finance (Mardle, Pascoe, & Tamiz, 2000). Compared to these conventional methods, multi-objective evolutionary algorithm (MOEA) seems more suitable to solve multi-objective problems because it searches a set of possible solutions simultaneously. Therefore, we can obtain a set of Pareto solutions rather than a special solution as in conventional methods in a run. These solutions are very important for the decision-maker to chose the optimal expansion process because some objectives are intangible and cannot be form using the conventional mathematical model. In addition, another reason which we use MOEA in this paper is that it can deal well with concave and discountinuous objective functions and Pareto frontiers.
On the other hand, the concept of fuzzy sets was proposed by Zadeh (Kim, Modkowitz, & Koksalan, 1965) to represent the uncertain situations or the subjective judgments. Using the membership function, we can measure the degrees of the uncertainty and deal with the fuzzy problems. Recently, the concept of fuzzy sets has been incorporate into the conventional statistical or mathematical programming methods to reflect the ambiguity and uncertainty in practice (Zimmermann, 1978, Tanaka and Lee, 1998, Kim et al., 1996).
In this paper, we propose the fuzzy multi-criteria competence set analysis. A numerical example is demonstrated to select the optimal fuzzy multi-criteria expansion process. Two criteria (cost and benefit function) with fuzzy numbers are used to reflect the ambiguity and uncertainty in practice. By employing MOEA, we can obtain Pareto solutions. On the basis of Pareto solutions, decision-makers can determine the finial optimal expansion process based on his preferences or subjective judgments.
The remainder of this paper is organized as follows. The expansion process of competence set analysis and the proposed method are discussed in Section 2. Multi-objective evolutionary algorithm is proposed in Section 3 to describe its ideas and procedures. A numerical example is used to demonstrate the proposed method in Section 4. Discussions are presented in the Section 5 and conclusions are in the last section.
Section snippets
Expansion process of competence set
The concept of competence set was proposed by Yu (1990) to resolve a particular decision problem by acquiring the necessity of ideas, information, skills, and knowledge. The contents of competence set analysis are to identify the true competence set, the decision-maker's competence set, and the efficient expansion path to make good decisions.
Among these issues, the method to optimally expand the existing competence set is especially highlighted. Several methods, such as the minimum spanning
Multi-objective evolutionary algorithm
Multi-objective evolutionary algorithm (MOEA) has been widely used since the 1990's to resolve the combinational problem in various areas such as scheduling (Murata, 1996), engineering (Fonseca & Fleming, 1998) and finance (Mardle et al., 2000). The concept of MOEA is based on the method of genetic algorithm (GA). GA was pioneered in 1975 by Holland, and its concept is to mimic the natural evolution of a population by allowing solutions to reproduce, create new solutions, and compete for
Numerical example
In this numerical example, we will demonstrate a fuzzy two-criterion (i.e. cost and benefit) expansion of competence sets. Let SK={x0}, T\SK={x1,x2,x3,x4,x5,x6,x7} and the fuzzy cost and the fuzzy benefit functions, which represent with interval values, are shown in Table 1, Table 2. Note that the symbol, M, denotes the infeasible route and will be treated as a minimum number in our fuzzy mathematical programming model. In addition, the membership of the cost and benefit functions are assumed
Discussions
Competence set analysis has been used for many applications, such as learning sequences for decision-makers (Hu, Chen, & Tzeng, 2002) and for consumer decision problems (Chen, 2001, Chen, 2002). However, these papers only consider the situation of using one criterion and the crisp function. In practice, decision-makers usually determine the optimal expansion process based on multi-criteria which may be conflicting with each other. Therefore, Pareto solutions should be derived for
Conclusions
In this paper, we extend the conventional competence set analysis to consider the situation of multi-criteria and fuzzy number. In order to obtain Pareto solutions efficiently, MOEA is employed here. A numerical example is used to demonstrate the procedures of the proposed method. On the basis of the results, we can conclude that the proposed method can provide a more flexible and diverse model.
References (33)
- et al.
A parametric approach to fuzzy linear programming
Fuzzy Sets and Systems
(1986) Using competence sets to analyze the consumer decision problem
European Journal of Operational Research
(2001)Expanding competence sets for the consumer decision problem
European Journal of Operational Research
(2002)- et al.
Mathematical programming with multiple objectives: A tutorial
Computing and Operational Research
(1980) - et al.
Fuzzy versus statistical linear regression
European Journal of Operational Research
(1996) - et al.
An investigation of genetic algorithms for the optimization of multiobjective fisheries bioeconomic models
International Transactions of Operations Research
(2000) An approach to multicriterion optimization problems for engineering design
Computer Methods in Applied Mechanics and Engineering
(1978)- et al.
A foundation for competence set analysis
Mathematical Social Sciences
(1990) Fuzzy Sets
Information Control
(1965)- et al.
Management models and industrial applications of linear programming
Management Science
(1957)
Evolutionary algorithms for solving multi-ojbective proglems
Handbook of genetic algorithms
Multi-objective optimization using evolutionary algorithms
A fast and elitist multiobjective genetic algorithm: NSGA-II
IEEE Transactions on Evolutionary Computation
Multiobjective optimization and multiple constraint handling with evolutionary algorithms. Part II. Application example
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Genetic algorithms and engineering optimization
Cited by (33)
Competence assessment as an expert system for human resource management: A mathematical approach
2017, Expert Systems with ApplicationsCitation Excerpt :Huan et al. used multi-objectives evolutionary algorithms to optimize multi-criteria expansion of competence sets. The main innovation, according to their paper, is consideration of the multi-criteria perspective of competence sets (Huang, Tzeng, & Ong, 2006). Draganidis and Mentzas reviewed 22 commercial CMSs and 18 Learning Management Systems (LMSs) and provided further recommendations for research and development in this area (Draganidis & Gregoris, 2006).
Competence-based performance model of multi-skilled workers
2016, Computers and Industrial EngineeringCitation Excerpt :Thus the competence, or rather the lack of a competence, can be expressed in quantitative terms. Most of the works on competence focus on modeling and optimization of the competence set expansion process (Feng and Yu, 1998; Shi and Yu, 1999; Li et al., 2000; Huang et al., 2006; Lin, 2006), or its applications in various domains: human resource management (Małachowski, 2011), learning management (Hu, Chen, & Tzeng, 2002; Durand, Belacel, & LaPlante, 2013), and consumer decision models (Chen, 2001; Chen et al., 2002; Hu, Hu, Chen, & Tzeng, 2004). Linear programming, or from time to time evolutionary algorithms, are applied for competence expansion problems (Huang et al., 2006; Lin, 2006).
Optimizing cut order planning in apparel production using evolutionary strategies
2013, Optimizing Decision Making in the Apparel Supply Chain Using Artificial Intelligence (AI): From Production to RetailOptimal expansion of competence sets with multilevel skills
2012, Computers and Industrial EngineeringCitation Excerpt :Lin (2006b) proposed a multiobjective and multistage fuzzy competence set model, which used hybrid genetic algorithms to measure the inter-stage and intra-stage costs of optimizing the expansion process effectively. Huang, and Ong (2006) proposed two criteria, cost functions and benefit functions with fuzzy numbers, to select the optimal expansion process. Hu, and Chen (2004) and Hu (2007) utilized fuzzy sets to describe the level of skills, and then proposed a data mining technique and a relationship-based method based on grey relations to discover compound skills and to determine learning costs, respectively.
Multicriteria variable selection for classification of production batches
2012, European Journal of Operational ResearchCitation Excerpt :The goal is to select a set of variables that minimizes the cost of measuring and collecting data. Over the last decade, approaches for multicriteria variable selection using accuracy, sensitivity, and specificity have been proposed for several real life classification problems, especially in text recognition, business decisions and security systems; see Rose-Pehrsson et al. (2000), Doan and Horiguchi (2004), Piramuthu (2004), Pendaraki et al. (2005), Huang et al. (2006), Pasiouras et al. (2007) and Aragonés-Beltrán et al. (2008). For a comprehensive review on multicriteria classification and decision making, see Zopounidis and Doumpos (2002), Steuer and Na (2003) and Sueyoshi (2006).
An interactive fuzzy satisficing method based on fractile criterion optimization for multiobjective stochastic integer programming problems
2010, Expert Systems with Applications