A study of order based genetic and evolutionary algorithms in combinatorial optimization problems

Abstract

In Genetic and Evolutionary Algorithms (GEAs) one is faced with a given number of parameters, whose possible values are coded in a binary alphabet. With Order Based Representations (OBRs) the genetic information is kept by the order of the genes and not by its value. The application of OBRs to the Traveling Salesman Problem (TSP) is a well known technique to the GEA community. In this work one intends to show that this coding scheme can be used as an indirect representation, where the chromosome is the input for the decoder. The behavior of the GEA's operators is compared under benchmarks taken from the Combinatorial Optimization arena.