Multiple Criteria Decision Making (MCDM) refers to making decisions in the presence of multiple, usually conflicting, criteria which must be traded off in almost every engineering, business or personal decision. In actual practice the classification of MCDM problem is well fitted to the facets of problem solving; MADM is for selection (evaluation), MODM is for design. In the real world problem, it is necessary to combine MADM results with MODM method in order to perform best selection and optimal design of organizational resources. But, the united framework to implementing selection and design phases in the help of decision support systems (DSS) has not been suggested.
Furthermore, most complex decisions in organizations are made by groups of people considering multiple criteria. As complexity of organizational decision making increases, the need for multiple attribute group decision making (MAGDM) and for working in group increases. Although many researchers are interested in developing methods and useful group decision support for MCDM, they have paid little attention to reduce considerable cognitive burdens on the decision makers and to integrate resource planning problem with MAGDM results. Therefore, the aim of this research is to synthesize the methodologies of MCDM with intelligent DSS to utilize their benefits. For this purpose, this dissertation suggested two integrated directions:
1. An intelligent group decision support system (GDSS) for MAGDM is developed - group members’ preferences of alternative are represented a range-type, interactive procedures are provided to help the group reach a consensus, and group’s dominance conditions with group’s utility range are applied.
2. Knowledge based decision support system for goal directed resource planning is developed - a suitable MODM method for designing a set of optimal alternatives, resources, is examined, and a guideline for modeling and system development is suggested.
It is difficult for group members participating in decision making process to articulate their preferences with cardinal values and they are willing to provide only incomplete information about parameter values. These can take the form of incomplete information. In MAGDM, we suggest interactive procedures such that represent their preferences with utility ranges obtained from linear programming models with incompletely specified information, find conflicting opinions among group members’ information, and suggest a direction toward reaching a group consensus. The dominance conditions based upon group’s utility range for the group are presented.
Based upon the interactive procedure, we have developed an intelligent GDSS, called utility Range based Interactive Group decision support System (RINGS). RINGS gets the incomplete information about alternatives’ preferences and attributes’ weights, and then obtains automatically group members’ and the group’s utility ranges. Also, RINGS suggests graphically the conflicting opinions found with only incomplete information. Thus RINGS is an intelligent system for the group decision making.
Using RINGS developed, we have applied to an information system level evaluation of S-group’s four affiliated companies. The evaluation of IS has been an important issue in management and control of computer based IS during the last three decades, and is considered as an MAGDM problem due to its several dimensions. In this research, the selected four experts evaluate IS level of each company and provide incomplete information about companies, instead of using questionnaire. The main reason of IS evaluation of affiliated companies of S-Group is to check up the IS state and develop a competitive power.
Also, This research examines the application of multiple objective linear goal programming in the military resource planning/design. A solution is generally not made in a single step but is part of a process using multiple criteria decision support systems. We recommend an approach using a rule-based post-optimal analysis for further interactions, to overcome these difficulties and yield a satisfactory solution. This system can also provide a pragmatic approach to rigorously analyzing programming issue with the results of MAGDM.
