DISEÑO DE TRANSMISIONES DE ENGRANAJES CILÍNDRICOS MEDIANTE OPTIMIZACIÓN GENÉTICA

Abstract

In this paper a genetic algorithm (GA)-based optimization procedure for the design of gear transmissions is presented. For gear design, simultaneous discrete (p.e. pitch) and continuous variables nonlinear related are used. However, most optimization methods are suited for continuous design variables. The approach presented uses Gas as a tool to achieve not only the optimal design, but also a near-optimal designs. First, the optimization problem is formulated. It must be multiobjective (maximum strength, minimum energetic losses, etc) and restricted. A mechanism to transform the constrained problem into unconstrained thought penalty functions is proposed. Recommendations on the objective function and penalty terms are proposed. Next a design variables coding and decoding method, as well the genetic operators of reproduction, crossover and mutation are presented. Finally, it is analyzed an example in which the developed genetic algorithm has been used, comparing the obtained results from a previous optimization. Keywords: Genetic Algorithm, optimization, transmission, gear