Q-learning and MCTS techniques for improving an algorithm to compute discrete vector fields on finite topological spaces

  1. Julián Cuevas-Rozo 12
  2. Jose Divasón 2
  3. Laureano Lambán 2
  4. Ana Romero 2
  1. 1 Universidad Nacional de Colombia
    info

    Universidad Nacional de Colombia

    Bogotá, Colombia

    ROR https://ror.org/059yx9a68

  2. 2 Universidad de La Rioja
    info

    Universidad de La Rioja

    Logroño, España

    ROR https://ror.org/0553yr311

Aktak:
9th International Symposium on Symbolic Computation in Software Science (SCSS 2021), short and work-in-progress papers
  1. Temur Kutsia (ed. lit.)

Argitaletxea: RISC Report Series

ISSN: 2791-4267

Argitalpen urtea: 2021

Orrialdeak: 6-10

Biltzarra: 9th International Symposium on Symbolic Computation in Software Science (SCSS 2021)

Mota: Biltzar ekarpena

GOOGLE SCHOLAR lock_openSarbide irekia editor
Gordailu instituzionala: lock_openSarbide irekia Editor

Laburpena

In this work we present an ongoing project on the improving of a previous symboliccomputation algorithm computing discrete vector fields on finite topological spaces. To this aim, we consider different strategies to choose each one of the possible vectors at each step of the algorithm and we apply some reinforcement learning techniques.