Genetic algorithms for the airport gate assignment: Linkage, representation and uniform crossover

Abstract

A successful implementation of Genetic Algorithms (GAs) largely relys on the degree of linkage of building blocks in chromosomes. This paper investigates a new matrix representaion in the design of GAs to tackle the Gate Assignment Problem (GAP) at airport terminals. In the GAs for the GAP, a chromosome needs to record the absolute positions of aircraft in the queues to gates, and the relative positions between aircraft are the useful linkage information. The proposed representation is especially effective to handle these linkages in the case of GAP. As a result, a powerful uniform crossover operator, free of feasibility problems, can be designed to identify, inherite and protect good linkages. To resolve the memory inefficiency problem caused by the matrix representation, a special representation transforming procedure is introduced in order to better trade off between computational efficiency and memory efficiency. Extensive comparative simulation studies illustrate the advantages of the proposed GA scheme. © 2008 Springer-Verlag Berlin Heidelberg.