Методика оптимізаційного синтезу пакувальних машин із застосуванням генетичного алгоритму

Abstract

Розглянуто методику оптимізаційного синтезу структури пакувального автомата за допомогою генетичного алгоритму та описано конструкцію згенеровної машини. Evolutionary search methods successfully combine the advantages of heuristic methods and methods of mathematical programming. The genotype of packaging machine is presented by the set of genes, each coding the particular functional module of a machine: · product feeding tube; · forming collar · longitudinal seaming module ; · drawing module;· cross-seaming module; · cutting module ;· mechanisms of fold formation;· mechanisms of forming the bottom and the top of a package. The value of each gene at the start is determined by pseudorandom number generator and may be changed by mutation operator during the genetic algorithm run. The structure of a machine is optimized by several criteria (e.g. versatility, complexity, cost, energy, etc.) that form a complex cost function evaluating each member of the pool durnig the genetic algorithm operation. The genetic algorithm ends up with bulding a histogram (fig. 1) presenting the genotype of the optimal result. The phenotype is obtained by decoding the value of each gene. The 3D result models of several genetic algorithm runs are presented on fig. 2-4.