Research On Multi-Objective Optimization Particle Swarm Algorithm And Its Application

Particle Swarm Optimization algorithm is developed in the past few years to resolve the issue of multi-objective optimization algorithm intelligent group. The algorithm is based on this assumption: the various groups of particles can get effective message from past experience and the experience of other particles .The experiments was found that for the majority of optimization problems this algorithm has many features such as a faster convergence rate , the less the need for parameters and so on. But it remains inadequate in the distribution and convergence of its solution set. Its applications are less. From some current research, this paper designed a hybrid improved Particle Swarm Algorithm from three phrases, and applied to the Nutrition Catering Computing Model. The work in the dissertation is shown as follows:1. Summing up the multi-objective optimization solutions of the traditional methods and evolutionary algorithm based on the solution algorithm, and focuses on the particle swarm algorithm and its multi-objective optimization in the field of the status quo.2. This paper designed a new multi-objective PSO to improve the algorithm from three phrases: (1) It joined tabu algorithm and Crowded mechanism in selecting of the optimal value of the best, so that it avoided getting into the optimizing locality earlier and improved the solution precision. (2) A solution of an advanced fitness function, which introduced the conception of feasible zone to treat with the restrained conditions. (3) It put forward the non-dominated set: a strategy about the random choice in label switching. It improved the structural efficiency of the non-dominated set and reduced the algorithm's time complexity.3. It designed a new nutrition Catering model which can meet all nutritional needs of the human body and the needs of the various type of the crowd. This model is species diversity and flexible.In the end, the new nutrition Catering model is designed and realized with the new multi-objective PSO according to the characteristics. By comparison with the traditional way to solve the model and experimental results, it proved feasibility and effectiveness, expanding a new multi-target particle swarm algorithm applications.