Research For Multi-Attribute Decision-Making Algorithm Based On Evidence Theory And Cloud Model

On the basis of researching evidence theory and cloud model, we apply them with reinforcement learning in decision-making algorithm, and build a real, complete, universal decision-making system. By analyzing reinforcement learning method applied on artificial intelligence robot game system, we also supply assistance for decision-maker to make decision.We can solve the coexistence of fuzziness and randomness by using multiple attribution evaluation method based on cloud model theory and evidence theory. Compared with traditional multiple attribute decision making, the proposed method can not only make multi-attribution decision for some object to be evaluated, but also sort and select the best one from the options. Thus on the basis of evaluation for decision making, we provide a new method for research about decision-making.We design and implement a multi-attribution decision-making system based on the method as mentioned above. First, we make the probably attack combinations mode of both enemies and us as a strategy set, and determine the pay-off-function by quantitative results for confrontation situation analysis about both enemies and us. Second, we establish the complete information static game model. By solving the equilibrium mixed strategy game model of Nash pick, and combining with combat experience. Finally, we form a mission decision-making method.In addition, in order to improve the accuracy of evaluation results, this article combines computer simulation for robot combat characteristics, which is applied to the multiple attribute decision making to solve a lot of experts evaluation collection problem.