Research And Application Of Agent Union Generation Problem Based On Quantum Particle Swarm

Coordination and cooperation on a large scale across multiple Agent plays a very important role in the complex system of MAS. Is one of the important cooperation between the Agent coalition mechanism, which the coalition generation is key problem of multi Agent systems. This article has been based on studies using quantum-behaved Particle Swarm Optimization algorithm for Agent Coalition generation problems, extended and improved the methods for solving the problem.First swarm intelligence algorithms for single-task Agent coalition vulnerable to excessive accumulation of particles in the process for the build, cause a decrease in diversity, or even local optimal presents solution strategy based on improved quantum-behaved Particle Swarm, drawing on real-coded genetic algorithm in the excessive accumulation of particles in the Cauchy variation of particle aggregation degree of decrease, thereby maintaining the diversity of the particles. And we using a variety of group and optimal particle migration policies in parallel to speed up convergence. Experiments show that this policy can quickly and efficiently solve Agent Alliance, in terms of operating efficiency than similar methods.Then, build problem on the basis of the study of single Agent coalition, coalition of research on multitask serial build problem. Multitasking coalition most of the traditional method is to maximize the total revenue of the League as a goal, but in a real environment is not only considered the cost-benefit and need to consider issues such as time constraints, costs and other factors. In needs overall considered multifaceted factors of situation, traditional of to maximize total proceeds for target of adapted value function cannot adapted requirements, paper design a new of solution strategy to achieved coalition maximize proceeds, and coalition cost and coalition remaining can efforts, three aspects of balance, and according to task similar sex using match strategy in task began joined prior knowledge to accelerated task solution speed, for coalition generated provides meet multiple target of optimization programme.Finally, on the basis of these studies, generate Agent union strategy application of agile supply chain partner selection problem, appropriate improvement of quantum-behaved Particle Swarm, applied to partner selection problems. Experiments show that the algorithm suitable for partner selection problem and got very good results.In this paper, on the basis of existing research to develop ideas, will only solve the single objective of Agent coalition program expanded to be able to generate multiple targets balance between global optimization program for the Union to provide more diversity to generate solution ideas, and closer the actual work needs. In the face of a complex and uncertain, unknown environment, multi-objective of this paper to solve the program can take into account the interests of all parties more, closer to the actual job requirements and expectations, can provide a more reasonable solution for the actual work.