Every living system has a set of goals or equilibrium points to pursue and maintain, respectively. For example, we try to hit a mosquito when it approaches us because we do not want to be bitten. Most of the time, we are unsuccessful in hitting the mosquito because she wants to survive. Humans have attempted to exterminate mosquitoes, both manually and chemically. However, mosquitoes still ubiquitously exist because they succeed in their goals of “perpetuation of their species.”
Mosquitoes and higher living organisms, including humans, have multi-goals in the lives. In this example, mosquitoes have two goals, as follows: (i) survival, and (ii) perpetuation of their species. Humans at the higher or highest end of evolution have developed a set of complex goals, including the following: (i) survival and security, (ii) perpetuation of the species, (iii) feeling of self-importance, (iv) social approval, (v) sensuous gratification, (vi) cognitive consistency and curiosity, and (vii) self-actualization (Yu, 1990, 1995, 2002, 2009). Therefore, multiple criteria decision making (MCDM) problems, as part of the problems that living systems must solve, are not unusual.
Literature of human history recorded MCDM and their dynamic changes.
However, the application of dynamic decision making problems into mathematical analysis began in the nineteenth century by economists and applied mathematicians, including Pareto, Edgeworth, Von Neumann, and Morgenstern (Yu, 1985; Zeleny, 1982). The rapid development of MCDM occurred by expanding single-criterion optimization to multiple-criteria optimization, considering decision dynamics, examining multiple decision makers in the decision process, and introducing uncertainty, and unknown and neural physiology in the complex system of decision making (Tzeng & Huang, 2011). In the last four decades, research on MCDM covered
a wide range of concepts, methodologies, and application areas, which resulted in a substantial number of studies (Dong et al., 2005; Peng et al.; Shi, 2001; Shi et al., 2007, 2009; Tzeng & Huang, 2011; Willenius et al., 2008; Yu, 1985). The challenging decision problems in changeable space have seldom been studied (Chen, 2010; Chen
& Yu, 2009; Chen, Huang, & Yu, 2012; Yu & Chen, 2010a, 2010b).
Non-trivial decision making can be characterized by various dimensions of parameters that involve a number of elements, such as decision alternatives, decision criteria, decision outcomes, decision preferences, and decision information inputs.
According to Yu (1990, 2002) and Yu and Chiang (2002), Non-trivial decision making also involves the following four environmental facets: decisions as a part of the behavior mechanism, stages of the decision process, the players involved, and unknowns in decision making. These parameters can interact with each other, and vary with time and situations, and with changes in the psychological states of the decision makers involved. Some of them are observable, whereas some are hidden and can be neglected. To demonstrate this occurrence, consider the following example.
Example 1: Dog food
To relieve concerns of pet owners regarding the problem of overweight dogs, a dog food company designed a special package of dog food that was both nutritious and could reduce the weight of dogs. The company wanted to know if this product can be popularized; therefore, a statistical test was conducted to ask pet owners if they wanted to buy this special package. The statistical testing market was positive, and the company started “mass production.” Its dog food supply fell short of meeting the overwhelming demand. Therefore, the company doubled its capacity. However, after
one to two months of excellent sales, the customers and wholesalers began to return the dog food package because the dogs did not like eating it.
Several parameters in human behavioral systems, such as goal setting, state evaluation, charge structure, attention allocation, information inputs, decision elements, and decision environmental facets are involved in Example 1. Some of these parameters are noticeable, whereas others are not. The dog food company observed the problem of overweight dogs, which created a high level of charge (a precursor of mental stress) to pet owners; they noticed the pain and frustration of pet owners. To make a correct decision, they conducted a statistical market test to determine if the new product was worth producing. The positive statistical testing result was a crucial (and observable) decision parameter (information inputs) that prompted the company of mass producing this dog food. However, they neglected the key players, the dogs, which was a critical parameter that determined the outcome of the decision. Their inability to recognize this hidden key parameter resulted in the failure of the dog food company.
The MCDM has evolved rapidly over the past five decades.
Figure 1. The evolution of MCDM.
Its evolution is depicted roughly in Figure 1. Although Figure 1 is Deterministic
MCDM
Probabilistic and Fuzzy MCDM
MCDM with Unknown (illusion,
wishful thinking)
Decision Making in Changeable Spaces
(1) (2) (3) (4)
self-explanatory, a brief explanation is provided, as follows:
According to Habitual Domains (HDs) theory, human behavior gradually stabilizes, although it is dynamic. Therefore, people have habitual concepts and ways of thinking, acting, judging, and responding (Yu, 1990, 1995, 2002, 2009). The MCDM, as a part of human behaviors, may reach a steady state and exhibit habitual patterns with time. Consequently, in mathematical programming or ordinary decision-making problems, people may unwittingly assume that the decision parameters (or variables) are well-known or deterministic. Models and techniques for this type of problem belong to deterministic MCDM (box (1) in Figure 1).
Subsequently, people become aware of the existence of decision parameters that are not deterministic. By assuming that these non-deterministic parameters vary with a certain shape of probabilistic distributions or fuzzy membership functions, researchers developed models of probabilistic and fuzzy MCDM (box (2) in Figure 1), such as Carlsson and Fuller (1996), Chiou and Tzeng (2002), Hsu et al. (2003), Lin et al.
(2010), Ana and Francisco (2011), and Opricovic (2011). However, certain parameters may be intangible in reality. We may be unaware of their existence without special effort, as demonstrated in Example 1. Even when they are noticed, their dimensions, ranges, and shapes may not be easily predetermined or assumed, as shown in box (3) of Figure 1. The perception of decision makers may be subject to the influence by their wishful thinking and illusions. All intangible or invisible parameters can be grouped under an umbrella of “unknown.” Decision models with these features, including contingency solution, are called MCDM with unknown (box (3) in Figure 1).
However, several real challenging decision problems involve invisible parameters, and the related parameters are also changeable as the situation and psychological states of mind of the decision maker changes. Discovering and controlling the change
of these parameters is often a vital part of the process to solve challenging decision problems. This type of problems is called decision making in changeable spaces (box (4) in Figure 1) (Yu & Chianglin, 2006; Yu & Chen, 2010a, 2010b, 2010c). The main difference between the problems of boxes (3) and (4) is that box (3) supplants all unspecified parameters into “unknown,” whereas box (4) actively searches for the unspecified parameters, which are treated as crucial variables for the solutions. Please refer to Appendix 1 for a more precise delineation of the main differences among the models of boxes (1) to (4).
Because most current MCDM studies (Dong et al., 2005; Peng et al.; Shi, 2001;
Shi et al., 2007, 2009; Tzeng & Huang, 2011; Willenius et al., 2008; Yu, 1985) are based on the assumption that the parameters are well-known or partially (fuzzily or probabilistically) known, their results cannot be applied effectively in the challenging decision problems in changeable spaces. Optimal solutions to the challenging problems in changeable spaces are usually found after these parameters are properly located, studied, searched, and restructured. The challenging decision problems of boxes (3) and (4) may also contain a number of parameters with the same properties of those in boxes (1) and (2). Once the crucial parameters of challenging problems (boxes (3) and (4) in Figure 1) are explored and located properly, this problem may become a fuzzy or deterministic problem (boxes (1) and (2) in Figure 1), and the models and techniques developed in boxes (1) and (2) can subsequently be used to sharpen and/or identify the “best” solution for the problems (see Chapter 3 for a detailed classification of a routine, fuzzy, and challenging problem).
According to Yu (1990, 2002), “Superior strategists find the best strategies by changing the relevant parameters, while ordinary strategists find the optimal solutions
within some fixed parameters.” Every living system has a set of goals or equilibrium points to pursue and maintain, respectively. We must interact with other people or organizations in our life processes. Theoretically, the interactions among people and/or organizations are the interactions of the HDs that they carry. The importance of knowing the HDs of ourselves, other people, and/or organizations cannot be overstated in solving nontrivial decision problems. HDs provides a set of unified concepts and techniques for us to sharpen our capacity to know ourselves, others, and environments, and to form optimal strategies for solving our problems (Yu, 2002).
These concepts1, in a similar manner to social comparison theory, identification spheres, and similarity effects, can be quantified for more specific use. Identification sphere is chosen because (i) it is an important phenomenon in human relation, an important part of human systems, and its pervasive influence on human behavior and decision making, including social networking, election, strategic alliance, marketing promotion, etc (Adler & Adler, 1987; Kreiner & Ashforth, 2004; Sluss & Ashforth, 2007; Yu, 2002), and (ii) by knowing a person’s identification spheres over different events and people/organizations, to a large degree, we could roughly know that person’s HDs. To illustrate this, chapters 6－7 will focus on the quantification and applications of identification spheres.
The main contributions of this dissertation can be listed as follows:
This dissertation explores the dynamic nature of decision making. By being unaware of the dynamic nature of the decision problems, decision makers may unwittingly fall into decision blinds, traps (Yu & Chianglin, 2006), and shocks, and make considerable mistakes. Understanding the behavioral dynamics and
1 Please see Yu (2002) for more concepts of HDs and it’s significant impacts on decision making.
HDs of ourselves and others can enable us to examine, search, and identify the optimal change of the relevant parameters to become superior strategists. A checklist of decision parameters is provided to help decision makers gain insight into the challenging problems and their possible solutions in changeable spaces.
Based on HD theory and competence set analysis (described in Chapter 3), a framework to solve challenging decision problems is provided, that is, Innovation Dynamics (Chen, 2010; Chen & Yu, 2009; Chen, Huang, & Yu, 2012;
Yu & Chen, 2010a, 2010b). It systematically describes processes of dynamic decision making in changeable spaces. By examining the operations of each link in innovation dynamics, decision makers can understand if each link is properly developed to continually upgrade their products/services and create maximal value by relieving the pains and frustrations of customers in the potential domains (PDs, one of the four sub-concepts of HDs, is described in Section 2.1).
This framework also indicates that each link must be properly examined and developed. Omitting any link can lead to substantial or disastrous mistakes.
The quantitative model of identification spheres based on HD theory was first suggested in several research areas, including psychology, political science, economics, and management science (Adler & Adler, 1987; Kreiner & Ashforth, 2004; Sluss & Ashforth, 2007). We defined and explained the concepts of identification function, degree of identification, identification matrices, identification spheres, and dis-identification spheres. The quantitative concepts were subsequently applied to formulate the mathematical models of election of a social group leader. The models considered both degree of identification and that of dis-identification. Different models lead to different results, which enrich our
thinking on elections. We also introduced mathematical programming models, and the use of identification matrices and thresholds of identification to identify an optimal combination of key members, to influence all targets in a social group.
The results are crucial for marketing promotion of a new product, service, or concept.
The remainder of this dissertation is divided into seven chapters. In Chapter 2, the relevant literature and concepts of HD and identification sphere are described, including their considerable effects on decision making.
As a vital application of HD, the concept of competence set and related decision analysis, including decision blinds, decision trap, and decision shocks are introduced in Chapter 3. This chapter contains a description of a checklist of decision parameters.
In Chapter 4, an anatomy of Innovation Dynamics, a systematic framework, is described. Innovation dynamics consists of a number of key components, which involve a number of decision parameters. By assessing the list of these potential challenging problems and decision parameters, decision makers may gain insight into challenging problems and their possible solutions in changeable spaces.
Three tool boxes of HD are provided in Chapter 5 to help decision makers identify and discover hidden parameters, and to make effective decisions in changeable spaces. Chapter 6 details the formation of identification and quantitative identification model, including identification function, degree of identification, identification matrices, identification spheres, and dis-identification spheres.
Applications of identification spheres are explained in Chapter 7. Section 7.1
provides the applications of the quantitative model to electing a group leader. In this approach, voters can express their degree of identification with candidates, as well their degree of dis-identification. Furthermore, we present various forms of election by using this new rule of election, which leads to various election results and enriches our thinking on elections. Section 7.2 presents the application of the introduced identification concepts to marketing promotion. Specifically, we introduce a mathematical programming model, and the use of identification matrices and thresholds of identification to identify an optimal combination of key members to influence all targeted people in a social group. To promote its products or services effectively to its targeted customers, corporations can first negotiate with this set of key members before launching new products or services. Finally, Chapter 8 offers a conclusion and suggestions for future research.