lpPSO - New optimization strategy inspired by PSO

Abstract

Given the large number of optimization problems that mankind faces, metaheuristics are very important strategies for the resolution of these problems. These strategies assess the suitability of the individuals, which represent solutions to the problem, a large number of times throughout the search for an optimal solution. When the assessment of an individual takes significant time or resources, the assessment of hundreds or thousands of individuals is a problem to be taken into consideration. In this paper, a strategy based on PSO that considerably reduces the number of individual assessments is presented, which is of great help for complex problems. The method proposed was compared with the classical version of PSO using classic functions in the space and a real case with a simulation model, and satisfactory results were obtained.