A Study On MCDM And Its Application In Ship Design

People have dealt with multiple criteria for as long as there were decision to be made. However, the concept did not receive its first "scientific" articulation until forties in this century, and the field of multiple criteria decision making(MCDM) is still a virgin land that need to develop even today. There is little effective analytic technique to many significance problems such as uncertain decision making, which is also one of the most popular topics in MCDM. In this dissertation, uncertain MCDM has been selected as major research work, and genetic algorithm being selected as means to solve certain and uncertain MCDM. As examples, the decision models and algorithm have been applied to ship design in this dissertation. The dissertation consists of three parts, including nine chapters.There is only one chapter in the first part. In this part, the present situation and development tendency of MCDM and its application(especially in ship design)are introduced. The concept of computational intelligence(CI) is outlined. The necessity of applying CI into MCDM is described is this part also.The second part is concerned with uncertain MCDM, including uncertain multiple attribute decision making(MADM), uncertain multiple objective decision making(MODM). And a kind of method is also put forward to calculate index weight of decision in this part. There are three chapters in this part. In chapter 2, A kind of order of uncertain attribute is presented first, and then uncertain MADM models is established, which including interval MADM, fuzzy MADM. The models are applied to program fast ships on a channel line, and some conclusions are drawn. Chapter 3 present uncertain MODM. Difference of various definitions of possibility is demonstrated first in chapter 3. Fuzzy evaluating principle is extended into fuzzy stochastic MODM, and possibility-reliability decision model ( PR model ) , possibility-reliability-satisfaction decision model(PRS model) are set up to uncertain MODM based on possibility of objective function, reliability of objective function and satisfaction of objective function. The decision models can be used widely in complex decisions that concerning with uncertain factors. Chapter 4 discusses a kind of method of interval statistics to determine the index weight of decision, which can treats the fuzzy and disagreement of specialist opinions freely, and analyze the evaluation levels of specialists.The third part is concerned with genetic algorithm (GA) and its application in MODM. Genetic algorithm is used to solve noninferiority set of certain MODM inthis part. The examples are demonstrated to analyze the efficiency of genetic algorithm for MODM. The genetic algorithm for MODM is applied to dimensions optimization of Chang jiang maritime affairs ship, 300TEU coastal container ship and the ship(fleet) type programming that passing through the locks of three gorges dam. Genetic algorithm is also designed for various uncertain MODM models on basis of stochastic simulation and fuzzy simulation. A struction of three-truss is selected as an example to demonstrate genetic algorithm for uncertain MODM models.Genetic algorithm is efficient for some MODM models, which are difficult to build or deal with by convention means. It's convenience for the people who dealing with decision to take uncertain factors into account directly through various uncertain MODM models. The high computability ensure practicability of the models, which can be used in wide fields such as pattern identification, association analysis % fuzzy reasoning, image identification. language processing, fault diagnosis . investment decision, modern management, system evaluation, artificial intelligence and expert system.