Multi-objective Monte-Carlo Tree Search
نویسندگان
چکیده
Concerned with multi-objective reinforcement learning (MORL), this paper presents MOMCTS, an extension of Monte-Carlo Tree Search to multi-objective sequential decision making. The known multi-objective indicator referred to as hyper-volume indicator is used to define an action selection criterion, replacing the UCB criterion in order to deal with multi-dimensional rewards. MO-MCTS is firstly compared with an existing MORL algorithm on the artificial Deep Sea Treasure problem. Then a scalability study of MOMCTS is made on the NP-hard problem of grid scheduling, showing that the performance of MO-MCTS matches the non RL-based state of the art albeit with a higher computational cost.
منابع مشابه
Cooperative Games with Monte Carlo Tree Search
Monte Carlo Tree Search approach with Pareto optimality and pocket algorithm is used to solve and optimize the multi-objective constraint-based staff scheduling problem. The proposed approach has a two-stage selection strategy and the experimental results show that the approach is able to produce solutions for cooperative games.
متن کاملProbabilistic Power Distribution Planning Using Multi-Objective Harmony Search Algorithm
In this paper, power distribution planning (PDP) considering distributed generators (DGs) is investigated as a dynamic multi-objective optimization problem. Moreover, Monte Carlo simulation (MCS) is applied to handle the uncertainty in electricity price and load demand. In the proposed model, investment and operation costs, losses and purchased power from the main grid are incorporated in the f...
متن کاملAn Intelligent Optimization Model for Multi-objective Order Allocation Planning
This paper presents a multi-objective order allocation planning problem with the consideration of various real-world production features. A novel hybrid intelligent optimization model, integrating a multi-objective memetic optimization process, a Monte Carlo simulation technique and a heuristic pruning technique, is proposed to handle this problem. Experiments based on industrial data are condu...
متن کاملEnhancements for Multi-Player Monte-Carlo Tree Search
Monte-Carlo Tree Search (MCTS) is becoming increasingly popular for playing multi-player games. In this paper we propose two enhancements for MCTS in multi-player games: (1) Progressive History and (2) Multi-Player Monte-Carlo Tree Search Solver (MP-MCTS-Solver). We analyze the performance of these enhancements in two different multi-player games: Focus and Chinese Checkers. Based on the experi...
متن کاملMonte-Carlo Tree Search: Applied to Domineering and Tantrix
................................................................................................................................................... i Chapter 1: Introduction ........................................................................................................................... 1 The Rules of Tantrix ...............................................................................
متن کامل