Based On Particle Swarm Optimization For Multi-objective Function

Optimization techniques is to apply technology to solve practical problems of the math. It is widely used, such as structural design, logistics and other areas, has played a great social and economic benefits. Optimization is the way of thinking has become an important class of applied mathematics, geometry, algebra, probability and computer science, system science, automation, there is close contact. At home and abroad show that under the same conditions, optimized to improve system efficiency of the right to technical processing, reduce energy consumption, the rational use of resources and effective way to improve the economy has a good effect, but also dealing with the object increases, this effect more pronounced. Particle swarm optimization (PSO: ParticleSwarmOptimization) algorithm is a new type of optimization technology, which comes from the idea of artificial life and evolutionary computation theory. Particle swarm optimization is through the particle to pursue their own to find the optimal solution and a whole group of the best solution to complete the optimization. In order to avoid particle swarm optimization algorithm to seek the optimal solution will be trapped into local optimization problems and to improve particle swarm optimization algorithm convergence rate between the So, we propose a particle swarm optimization algorithm to increase the probability update method. Pairs of unconstrained and constrained optimization problems were designed based on particle swarm optimization algorithm for solving the different methods and test functions, and the particle swarm optimization algorithm for solving multi-objective optimization problems of further study. Experiments show that the improved particle swarm optimization algorithm for optimization problems play a certain validity. Multi-objective function optimization is the fundamental problems facing humanity, life is often their own experience or intuition to rely on engineering solutions, often single-objective problem is a multi-objective problem. However, due to a variety of complex objective nature of the problem itself, such a solution does not reflect the objective nature of the problem, can not meet the normal requirements. Particle swarm intelligence heuristic clustering algorithm belonging to the probability of it from the bird migration and feeding activities, extract the corresponding model, a large number of functions have been used to solve various optimization problems. May be used to solve multi-objective function optimization problem with a certain degree of difficulty: first, the multi-objective problem itself is difficult to determine an appropriate, and not easy to easy to resolve; In addition, PSO algorithm for multi-objective optimization, there is no solid basis for mathematical theory of genetic algorithms different, is still at the early stages of research. This is the theoretical basis for the study is still relatively poor, the researchers are also mechanisms for particle swarm can not be the appropriate mathematical explanation. However, the advantages of the algorithm is simple and easy to implement, while the background has a profound intelligence work, both for scientific research, particularly suitable for engineering application of appropriate benefits, opening up the application of PSO algorithm is a new area of the valuable work because the PSO algorithm lies in the vitality of engineering applications. The main contents of this article: first, mainly introduces the multi-objective optimization theory and particle swarm optimization algorithm the origin and development to discuss the computational intelligence techniques for solving multi-objective function optimization of the status quo and development trend, highlighting the PSO algorithm in multi-objective optimization Research in the field of study of the paper goes on to identify the content and direction. Second, a multi-objective optimization focuses on the theory of knowledge-based, compiled in the traditional multi-objective optimization methods and classical solution methods of some of. Third, the particle swarm algorithm is a detailed analysis and synthesis, particle swarm optimization algorithm is given the basic principles, algorithms, processes, and parameter settings, use the Java language to implement the basic PSO algorithm. Finally by comparing the basic idea of genetic algorithms and implementation steps of genetic algorithms and particle swarm optimization algorithm discussed the link between the particle swarm optimization algorithm to illustrate the multi-objective function optimization to achieve faster convergence on the optimal solution. Fourth, based on the Java language has developed a particle swarm optimization algorithm simulation procedures, including the Rosenbrock function; Rastrigrin function; Sphere function; Griewank function; Ackley function, and many other test functions, by setting the number of iterations, population size, inertia weight, learning factor and other relevant parameters, the program can be a variety of test functions for the different dimensions of the experiment, movement panel to observe the movement of particles search target, fitness panel shows the optimal value curve, the average curve and difference curve. Through a function test and analysis of the particle swarm optimization algorithm a good search performance. Fifth, the main application the first two chapters has concluded that concrete realization based on particle swarm algorithm for multi-objective optimization problems in-depth study and assessment of the optimal solution using the selected algorithm is used for multi-objective optimization problems the optimal non-inferior solution set of the search. Through experiments on several test functions to prove its validity. Finally, the full text of the conclusion that the experimental data show that in order to assess the optimization problem more difficult to select the algorithm can accurately shows the Pareto curve shape and most individuals falling into non-inferior set of the optimal solution algorithm for the optimal solution to the effective sex. This article is the work of verification. For the algorithm to further improve and solve objective optimization .