Browsing by Author "Castillo, Mauricio"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    A New Learnheuristic: Binary SARSA - Sine Cosine Algorithm (BS-SCA)
    (Springer Science and Business Media Deutschland GmbH, 2022) Becerra-Rozas, Marcelo; Lemus Romani, José Isaac; Crawford, Broderick; Soto, Ricardo; Cisternas Caneo, Felipe; Trujillo Embry, Andrés; Arnao Molina, Máximo; Tapia, Diego; Castillo, Mauricio; Rubio, José Miguel
    This paper proposes a novel learnheuristic called Binary SARSA - Sine Cosine Algorithm (BS-SCA) for solving combinatorial problems. The BS-SCA is a binary version of Sine Cosine Algorithm (SCA) using SARSA to select a binarization operator. This operator is required due SCA was created to work in continuous domains. The performance of BS-SCA is benchmarked with a Q-learning version of the learnheuristic. The problem tested was the Set Covering Problem and the results show the superiority of our proposal.
  • No Thumbnail Available
    Item
    A Novel Learning-Based Binarization Scheme Selector for Swarm Algorithms Solving Combinatorial Problems
    (2021) Lemus-Romani, Jose; Becerra-Rozas, Marcelo; Crawford, Broderick; Soto, Ricardo; Cisternas-Caneo, Felipe; Vega, Emanuel; Castillo, Mauricio; Tapia, Diego; Astorga, Gino; Palma, Wenceslao; Castro, Carlos; Garcia, Jose
    Currently, industry is undergoing an exponential increase in binary-based combinatorial problems. In this regard, metaheuristics have been a common trend in the field in order to design approaches to successfully solve them. Thus, a well-known strategy includes the employment of continuous swarm-based algorithms transformed to perform in binary environments. In this work, we propose a hybrid approach that contains discrete smartly adapted population-based strategies to efficiently tackle binary-based problems. The proposed approach employs a reinforcement learning technique, known as SARSA (State-Action-Reward-State-Action), in order to utilize knowledge based on the run time. In order to test the viability and competitiveness of our proposal, we compare discrete state-of-the-art algorithms smartly assisted by SARSA. Finally, we illustrate interesting results where the proposed hybrid outperforms other approaches, thus, providing a novel option to tackle these types of problems in industry.
  • No Thumbnail Available
    Item
    Reinforcement Learning Based Whale Optimizer
    (Springer International, 2021) Becerra Rozas, Marcelo; Lemus Romani, José Isaac; Crawford, Broderick; Soto, Ricardo; Cisternas Caneo, Felipe; Embry, Andres Trujillo; Molina, Maximo Arnao; Tapia, Diego; Castillo, Mauricio; Misra, Sanjay; Rubio, Jose Miguel
    This work proposes a Reinforcement Learning based optimizer integrating SARSA and Whale Optimization Algorithm. SARSA determines the binarization operator required during the metaheuristic process. The hybrid instance is applied to solve benchmarks of the Set Covering Problem and it is compared with a Q-learning version, showing good results in terms of fitness, specifically, SARSA beats its Q-Learning version in 44 out of 45 instances evaluated. It is worth mentioning that the only instance where it does not win is a tie. Finally, thanks to graphs presented in our results analysis we can observe that not only does it obtain good results, it also obtains a correct exploration and exploitation balance as presented in the referenced literature.