Experimental research through randomness points as metaheuristics suboptimal local responses improvement
Metaheuristic algorithms are widely used in the optimization of problems in different areas. Several studies have, for example, applied this method to the optimization of truck logistics in open pit mining. This research approaches an experimental analysisof the GRASP* metaheuristic through the vari...
| Authors: | , , , , , , |
|---|---|
| Format: | article |
| Status: | Published version |
| Publication Date: | 2022 |
| Country: | Brasil |
| Institution: | Associação Brasileira de Engenharia de Produção (ABEPRO) |
| Repository: | Revista Produção Online |
| Language: | Portuguese |
| OAI Identifier: | oai:ojs.www.producaoonline.org.br:article/4398 |
| Online Access: | https://www.producaoonline.org.br/rpo/article/view/4398 |
| Access Level: | Open access |
| Keyword: | Metaheurística GRASP GRASP* Roteamento Minas a Céu Aberto Metaheuristics Routing Open-pit mines |
| Summary: | Metaheuristic algorithms are widely used in the optimization of problems in different areas. Several studies have, for example, applied this method to the optimization of truck logistics in open pit mining. This research approaches an experimental analysisof the GRASP* metaheuristic through the variationof the randomness point, with metrics not yet explored in the literature, in order to verify the performance of the algorithm in relation to suboptimal solutions. After the analysis of the algorithmconvergencywith the changes on the randomness points, a study of its performance in relation to the amount of processing cycles was performed. Databases alreadyevaluated in other studies, added to 10 other reference databases present in the literature, were employed during the exploratory analysis of the GRASP* method. In addition, the results obtained by the GRASP* algorithm were compared with the NN* constructive heuristic. The results of this study demonstrate that the changes applied to the GRASP* method provided gains of more than 24% in performance for given values of randomness point and gains of more than 10% with varying numbers of cycles. Such a framework can be implemented for the optimization of logistical strategies that can drive million-dollarbusinesses, such asopen pit mining. |
|---|