The present research is focused on the development of a facility layout problem (FLP) using genetic approach for robust optimization (i.e., GA) and analysis of Variance ANOVA used to identify critical parameters of the genetic algorithm. The main advantage of such type of algorithms with respect to the other artificial intelligent approaches is represented by the possibility to simulate the natural evolution of a survival fitness function and, in the specific case, the minimization of material handling costs for different located machines. Hard constraints (e.g., machine overall dimensions and production cycle) have been taken into consideration. The implemented algorithm uses a string encoding, a partially matched crossover (i.e., PMC) to generate the offspring and a repair method to modify infeasible solutions. The repair method analyzes the solution, identifies the plants in which the constraints are not observed and places a random machine into another plant. The input data are production cycles, areas of a plant, machines and storehouse. To investigate the effects of individual parameters of GAs, a design of experiments (DOE) was employed. Early analysis of factors (i.e. dimension of population, fraction of elitism and fraction of eliminated genes during generations) was conducted with a full factorial plan with two levels for each factor and four repetitions. A trade-off of the population dimension and maximum number of generations allows to achieve good results in short computation time. The results are presented in terms of reduction of external transport costs.

An Automated Procedure for Machines Distribution among Plants based on Genetic Algorithm Approach

LAMBIASE, FRANCESCO;
2009-01-01

Abstract

The present research is focused on the development of a facility layout problem (FLP) using genetic approach for robust optimization (i.e., GA) and analysis of Variance ANOVA used to identify critical parameters of the genetic algorithm. The main advantage of such type of algorithms with respect to the other artificial intelligent approaches is represented by the possibility to simulate the natural evolution of a survival fitness function and, in the specific case, the minimization of material handling costs for different located machines. Hard constraints (e.g., machine overall dimensions and production cycle) have been taken into consideration. The implemented algorithm uses a string encoding, a partially matched crossover (i.e., PMC) to generate the offspring and a repair method to modify infeasible solutions. The repair method analyzes the solution, identifies the plants in which the constraints are not observed and places a random machine into another plant. The input data are production cycles, areas of a plant, machines and storehouse. To investigate the effects of individual parameters of GAs, a design of experiments (DOE) was employed. Early analysis of factors (i.e. dimension of population, fraction of elitism and fraction of eliminated genes during generations) was conducted with a full factorial plan with two levels for each factor and four repetitions. A trade-off of the population dimension and maximum number of generations allows to achieve good results in short computation time. The results are presented in terms of reduction of external transport costs.
978-88-95028-38-5
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11697/32183
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact