Evolutionary Based on Selfish and Altruism Strategies - An Approach to Path Planning Problems

This paper presents a novel hybrid optimization approach based on a genetic algorithm that combines selfish gene and altruism view of evolution. The purpose of the present research is to develop a new optimization approach to solve path-planning problems, particularly to be used in robot trajectories planning. A brief discussion about selfish versus altruism is made in the perspective of genes, its integration in the chromosome (individuals) and the forces involved in the evolution process of genes, individuals and populations. The SAGA (Selfish-Altruist Genetic Algorithm) is the generalization of the Genetic Algorithms (GA), where the basic variables are the genes (characters or words) as non-autonomous entities, grouped in a Chromosome structure. The proposed hybrid approach was applied to a path-planning problem, in a continuous search space, to show its effectiveness in complex and interdependent sub-paths and evolution processes. Genes-centred evolution improved local sub-paths as sub-processes in a Chromosome-centred evolution and resulted in improved global planning trajectories when compared with standard genetic algorithm.

I. Introduction

The theory of evolution is a new light in human thought and it has today transposition to computational models. Successive Evolutionary Algorithms (EA) and artificial life have been presented during the last 4 decades through computational simulations to solve complex engineering problems [1], particularly in optimization. Genetic algorithms have become one of the most famous EA algorithms since the pioneering work of John Henry Holland at 1975 [2]. They are based on Darwin’s natural selection theory and were extended today to the automatic evolution of computer programs [1]. These evolutionary algorithms are now used to solve complex multi-dimensional problems more efficiently than other traditional methods, which do not own appropriate features that restrict their general use [3][4].