Research And Application In Group Decision Based On Social Chioce Theory

This thesis is researched in the group decision problems based on social choice theory. From a theoretical point of view, logic and social choice theory are applied in this thesis for solving the group decision problems. Group decision result makes a great influence to members’ interests in the group. Therefore, group decision should be finished by all the members of the group. Each member should freely and honestly express his/ her opinions of the issues in the collective decision procedure, the collective opinion outght to aggregate all these opinions or preferences to reach a collective opinion or preference. So the inconsistent collective judgment set and backstage manipulation can not meet the will of the most of voters and influence the fair and just of the group decision. There are some solutions for these two main problems in group decision procedure in this thesis. From a practical point of view, to deal with the social choice methods as a aggregation mechanism is to sovle the cold-start problem in recommender systems, and find a new solution for cold-start problem, then it combins the logic and aritificial intelligence to be a new research point and it has a very bright future for practice.This work analyzed inconsistent collective judgment set and backstage manipulation bsed on social choice theory, and it is utilized the judgment agggragation model and preference aggregation model to be aggregation mechanism for solving the cold-start problem in recommender systems. The main research points in this work are presented in the following part:Firstly, solutions for inconsistent collective judgment set based on judgment aggregation model. From research in the inconsistent collective judgment set, it begins from the discursive dilemma, and applies the premise-based and conclusion-based methods, methods based on the majoritarian judgment set, methods based on the weighted majoritarian judgment set, and methods based on removal some individual judgment set to solve the inconsistent collective judgment set, and in order to eliminate the inconsistent judgment set to obtain the L-consitent collective judgment as the group decision making result.Secondly, towards an analysis to the backstage manipulation in group decision procedure nased on judgment aggregation model. According to the literatures related to backstage manipulation, we can see that the manipulation for judgment aggregation mechanism is one of the worst influence backstage manipulation ways. In order to prevent this kind of backstage manipulation happened; in this paper we design a method to defense the manipulation of judgment aggregation mechanism, which requires all individuals vote for the constraint first, we could get the opinions about the judgment aggregation mechanism of the voters, and then it can be obtain the collective judgment aggregation mechanism according to majoritartian aggregation rule. The social choice theory has proved that every aggregation method to achieve collective rationality based on voting has the possibility to be manipulated. How to control the manipulation and equitably obtain the group decision is a difficult problem in the aggregation procedure. This work analyzes the backstage manipulation in the group decision process based on judgment aggregation model, and we design the method to defense the manipulation of judgment aggregation mechanism according to the courses of backstage manipulation. It makes all the voters decide which judgment aggregation method applies to aggregate the individual judgment sets and obtain the collective judgment set. It analyze and find the solution for these two serious phenomena which influence the justice and equality of the group decision based on logic in order to lay a theoretical foundation for the development and application of multi-agent decision system.Moreover, a solution to the cold-start problem in recommender systems based on social choice theory. We focused on models of preference aggregation and judgment aggregation; specifically, by using the judgment aggregation model to solve the cold-start problem, which is a novel approach. It is easy to get the collective judgment set in the public interest. I utilized the linear Borda rule and Copeland rule based on preference aggregation model and Maxweight sub-agenda rule based on judgment aggregation model. This solution is combined social choice theory and logic, and treated the aggregation models in social choice theory as aggregation functions in recommender systems to solve the cold-start problem. The two contributions in this solution: on one hand, it is important to find a new solution for the cold-start problem, and it is not a classical methods and solutions in recommender systems; on the other hand, it gives a new idea which is combined social choice theory and artificial intelligence, and is developed the application of the social choice theory, meanwhile, it can give the new solution for the recommender systems.Group decision problem is analyzed based on models of social choice theory, which is a principle direction recently. Arrow Impossibility Theorem makes the democratic dream to be vanished, and this type of impossibility theorems spreads rapidly, which makes people deeply recogonise that the fair and just group decision cannot be realized. However, these kind of pathetic theorems have no help to the real group decision making procedure. If we give up collective decision because of the impossibility theorems, we should not stop eating for fear of choking. In this work we utilize the judgment aggregation model to solve the inconsistent collective judgment set and backstage manipulation problems in order to give some reasons for the real decision making procedure. We explore the reasons which lead to these two phenomenon and find some solutions for the inconsistent judgment set and the defence way for the manipulation, which can find a way to realize the fair and just group decision result. The research of theory is for the practical application. A new situation in modern social choice theory of multi-subject cross study has been formed. In this work we ustilize models of social choice theory to solve the cold-start problem in recommender systems, which is a new exploration, there are two main contributions: on the one hand, it expands the application range of social choice theory; on the other hand, it provides a new solution to the problems in the field of artificial intelligence.