The Artifical Physics Algorithm And Research On Its Application

With the development of the industry and the deepening of scientific research,the traditional algorithm seems to be limited in some complex problems.At the same time,swarm intelligence algorithm arises at the historic moment.As a member of new swarm intelligence algorithm,Artificial Physics Optimization algorithm has gotten the attention of scholars because of its advantage in principle,parameters and encoding.Scholars pay close attention to it only a short time,so that the theory is not perfect.In particular,there is less research on constraint optimization(Single target,MOP and interval MOP).In view of the above problems,the following is carried out in this paper.We have studied the basic principle and the insufficiency of APO,and found it's shortages including concentrated distribution initialization process,falling into local optimum problems into the late particles.In consideration of that,we have adopted some improvement mechanisms and then applied it to the solution to parameter identification of Wiener system and 31 provincial capital city China path optimization of complex functions;In the particle,we introduced the mathematical model of single objective constraint optimization problem and applied the filter technology to optimize this problem is other than the previous penalty function method.Numerical experiments are carried out to verify the feasibility of the algorithm and try to apply the algorithm to the optimization of actual engineering parameters.The multi-objective constrained optimization problem(MOP)mathematical model was defined and proposed two kinds of improved algorithm based on APO algorithm were proposed: In FMLAPO algorithm,a constrained violation function is used to evaluate the feasible domain particle quality,and limits on the Pareto front is not feasible solutions from the force rules using the random weighting method,which transformed MOP into Single Objective Problem.On the other hand,we proposed FMSAPO algorithm,in which the maximum entropy function method is used to transform the MOP problem into a single objective optimization,and then the filter technique is used to deal with the constraints.The proposed algorithms were used in the optimization of the constrained MOP problem to test its effectiveness.In the interval multi-objective constrained optimization problem(IMOP),we tried to transformed MOP optimization into the single objective optimization interval(ILPM)with fuzzy weighted rules;the interval mathematical theory provided determined-(single)-optimization model for us to discuss the upper and lower bounds of the interval.Given the number of the constraint conditions after transformation in the model more,and filter technology of constraints and strong processing ability,so using the FMAPO algorithm for solving it.