TY - CONF
AU - Marrero, A.
AU - Segredo, E.
AU - Leon, C.
T1 - A parallel genetic algorithm to speed up the resolution of the algorithm selection problem
LA - eng
PY - 2021
SP - 1978
EP - 1981
T2 - GECCO 2021 Companion - Proceedings of the 2021 Genetic and Evolutionary Computation Conference Companion
SN - 9781450383516
PB - Association for Computing Machinery, Inc
AB - Deciding which optimisation technique to use for solving a particular optimisation problem is an arduous task that has been faced in the field of optimisation for decades. The above problem is known as the Algorithm Selection Problem (ASP). The optimisation techniques considered in previous works have been, mainly, approaches that can be executed rapidly. However, considering more sophisticated optimisation approaches for solving the ASP, such as Evolutionary Algorithms, drastically increases the computational cost. We are interested in solving the ASP by considering different configurations of a Genetic Algorithm (GA) applied to the well-known 0/1 Knapsack Problem (KNP). This involves the execution of a significant number of configurations of the GA, in order to evaluate their performance, when applied to a wide range of instances with different features of the KNP, which is a computationally expensive task. Therefore, the main aim of the current work is to provide, as first step for solving the ASP, an efficient parallel GA, which is able to attain competitive results, in terms of the optimal objective value, in a short amount of time. Computational results show that our approach is able to scale efficiently and considerably reduces the average elapsed time for solving KNP instances.
DO - 10.1145/3449726.3463160
UR - https://portalciencia.ull.es/documentos/610dd3e20f233b05011f6e95
DP - Dialnet - Portal de la InvestigaciÃ³n
ER -