Forgetful Swarm Optimization for Astronomical Observation Scheduling

In this paper, we propose a novel metaheuristic algorithm called Forgetful Swarm Optimization(FSO) for Astronomical Observation Scheduling (AOS), a type of combinatorial optimization problemdefined by the tasks and constraints assigned to the telescopes and other devices involved in astrophysicalres...

ver descrição completa

Detalhes bibliográficos
Autores: Nakhjiri, Nariman, Salamó Llorente, Maria, Sànchez i Marrè, Miquel, 1964-, Blum, Christian, Morales, Juan Carlos
Tipo de documento: artigo
Estado:Versão publicada
Data de publicação:2024
País:España
Recursos:Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya)
Repositório:Recercat. Dipósit de la Recerca de Catalunya
OAI Identifier:oai:recercat.cat:2445/217730
Acesso em linha:https://hdl.handle.net/2445/217730
Access Level:Acceso aberto
Palavra-chave:Intel·ligència artificial
Aprenentatge automàtic
Algorismes computacionals
Artificial intelligence
Machine learning
Computer algorithms
Descrição
Resumo:In this paper, we propose a novel metaheuristic algorithm called Forgetful Swarm Optimization(FSO) for Astronomical Observation Scheduling (AOS), a type of combinatorial optimization problemdefined by the tasks and constraints assigned to the telescopes and other devices involved in astrophysicalresearch. FSO combines local optimization, Destroy and Repair, and Swarm Intelligence methodologies tocreate a flexible and scalable global optimization algorithm to handle the challenges of AOS. The proposalis adapted to the well-justified scenarios of the Ariel Space Mission problem, a particular example of AOS,and compared with previous algorithms that are applied to it including an Evolutionary Algorithm (EA),an Iterated Local Search (ILS), a multi-start metaheuristic, a Tabu Search, and a Hill-Climbing greedyalgorithm. The experimental evaluation demonstrates that FSO consistently outperforms other algorithmsin objective completeness, up to 8.4% on average, for all instances of the problem regardless of dimensionsand complexity. Additionally, it has significantly less computational cost than ILS and the base models of aglobal optimization algorithm such as EA.