ISA: a hybridization between iterated local search and simulated annealing for multiple-runway aircraft landing problem

Abstract

This paper presents an efficient method for aircraft landing problem (ALP) based on a mechanism that hybridizes the iterated local search (ILS) and simulated annealing (SA) algorithms. ALP is handled by scheduling each incoming aircraft to land on a runway in accordance with a predefined landing time frame. The main objective to address is to find a feasible aircraft scheduling solution within the range of target time. The proposed hybridization method complements the advantages of both ILS and SA in a single optimization framework, referred to as iterated simulated annealing (ISA). The optimization framework of ISA has two main loops: an inner loop and an outer loop. In the inner loop, SA is utilized through a cooling schedule and a relaxing acceptance strategy to allow ISA to escape the local optima. In the outer loop, the restart mechanism and perturbation operation of the standard ILS are used to empower ISA to broadly navigate different search space regions. Extensive evaluation experiments were conducted on thirteen small- and large-sized ALP instances to assess the effectiveness and solution quality of ISA. The obtained results confirm that the proposed ISA method considerably performs better than other state-of-the-art methods in which ISA is capable of reaching new best results in 4 out of 24 large-sized problem instances as well as obtaining the best results in all small-sized instances. Evaluation on large-sized instances confirms a high degree of performance. As a new line of research, ISA is an effective method for ALP which can be further investigated for other combinatorial optimization problems.