Search

This paper analyzes the effect of different configurations of initial conditions for the optimization algorithm based on particle swarms with vorticity behavior. The proposed algorithm combines the gradient-based search and particle swarm behavior, thus, this algorithm can be affected by the initial conditions for the particles. Finally, a 2D test function is used to observe the characteristics of the algorithm. [ABSTRACT FROM AUTHOR]

Evolution Strategies is a bio-inspired, robust, and efficient technique for solving optimization problems where the solution space is unrestricted. However, this assumption is unreal in many cases because the solution space is limited by complex boundaries in the form of linear and non-linear restrictions. In this paper, a modification of the original algorithm of Evolution Strategies for optimizing problems where the solution space is bounded using complex restrictions is proposed. The proposed method is based on the use of a penalization function which is zero inside of the feasible region and equal to the maximum value inside of the feasible region when an unfeasible point is considered. The proposed approach is proved using six benchmark problems. In all cases, our approach found an optimal point equal or lower than the values reported in the literature. [ABSTRACT FROM AUTHOR]

This paper presents an heuristics for solving p-median problems. The main idea is the execution of a preprocess routine whose goal is to reduce the number of variables facilitating its computacional resolution. This preprocessing is essentially based in the determination of a maximum distance that enables you to discriminate if two any nodes are served for the same facility. If the distance between these nodes is greater than the maximum distance, then these nodes will not be served for the same facility. As a consequence, variables need to be fixed to zero in order to reduce the size of the problem. [ABSTRACT FROM AUTHOR]