Optimization Algorithm, Based On Feasibility Rules The Mimicry Physics Constraints

Constrained optimization problem exists in the field of science and engineering. Therefore, it has been a hot topic of scholarly research. Artificial Physics Optimization (APO) is a novel stochastic optimization algorithm. It has been successful in solving global optimization problem.A feasibility-based rule is a very simple constraint processing technology, it has not additional parameters settings, and its implementation is simple. Artificial Physics Optimization Algorithm with a feasibility-based rule for constrained optimization problems framework is constructed.Introducing the feasibility-based rule, the individual is divided into feasible individual and infeasible individual. A feasibility-based rule is mainly studies between infeasible individual and feasible individual, between in two infeasible individuals'force rule. Three different feasibility-based rules are constructed in this paper, Through Theoretical analysis and simulation experiments show that Artificial Physics Optimization Algorithm with three different feasibility-based rules for constrained optimization problems are feasibility and effectiveness.In Artificial Physics Optimization Algorithm, the mass of feasibility individual is a user-defined function of the value of an objective function, and the mass of infeasibility individual is a user-defined function of the violate value function. Different the mass function of individual affect Artificial Physics Optimization Algorithm solving constraint optimization problems'performance. Through based on properties and characteristics of constraint optimization problems and analysis of the characteristics of mass function, according to linear, convex curve, concave curve constructed different types of mass function. The simulations results show that concave curve mass function's algorithm have better performance.