Year: 2025
Type: article
Abstract: ABSTRACT In the field of optimisation, the accurate configuration of parameters in metaheuristic algorithms is a critical yet often arduous task that significantly impacts the efficiency and efficacy of the search process. This study was motivated by the need to address the inefficiencies and limitations associated with conventional methods of parameter configuration, which typically involve manual, trial‐and‐error approaches. These traditional methods can lead to suboptimal performance and increased computational overhead. To tackle these challenges, this study introduces a novel adaptive parameter control strategy for parallel island‐based metaheuristics, with a particular emphasis on the ant colony optimisation (ACO) algorithm. Our research process involved extensive experimentation to evaluate the effectiveness of this adaptive strategy. We conducted a series of tests to enable real‐time adjustment of key parameters based on the performance of ACO colonies, thereby enhancing both exploration and exploitation capabilities. The results indicate that the adaptive strategy consistently outperforms offline manual and automated tuning configurations, particularly in larger and more complex problem instances, providing a more efficient solution for parameter optimisation in metaheuristics. These findings highlight the potential of dynamic parameter control to reduce dependency on expert knowledge and manual tuning while improving algorithmic performance.
Source: Expert Systems
DOI: 10.1111/exsy.70061
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,600
Awards: G1105844351, G1175115578, G1771795543, G2262748287, G294895050, G300979063, G3480869486, G410571670, G425668148, G451917667, G4637294384, G464105488, G4681143804, G491712853, G5080475149, G523076202, G5539308778, G5635532810, G593054264, G5967599077, G6033112984, G6233745249, G6731524417, G696953080, G7084143925, G7252005220, G7395345766, G7805626310, G7863371703, G8497050386, G8667377315
OpenAlex ID: W4409858582