Improved Monte-Carlo Search
نویسندگان
چکیده
Monte-Carlo search has been successful in many non-deterministic games, and recently in deterministic games with high branching factor. One of the drawbacks of the current approaches is that even if the iterative process would last for a very long time, the selected move does not necessarily converge to a game-theoretic optimal one. In this paper we introduce a new algorithm, UCT, which extends a bandit algorithm for Monte-Carlo search. It is proven that the probability that the algorithm selects the correct move converges to 1. Moreover it is shown empirically that the algorithm converges rather fast even in comparison with alpha-beta search. Experiments in Amazons and Clobber indicate that the UCT algorithm outperforms considerably a plain Monte-Carlo version, and it is competitive against alpha-beta based game programs.
منابع مشابه
A Monte Carlo-Based Search Strategy for Dimensionality Reduction in Performance Tuning Parameters
Redundant and irrelevant features in high dimensional data increase the complexity in underlying mathematical models. It is necessary to conduct pre-processing steps that search for the most relevant features in order to reduce the dimensionality of the data. This study made use of a meta-heuristic search approach which uses lightweight random simulations to balance between the exploitation of ...
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 کاملBiasing Monte-Carlo Simulations through RAVE Values
The Monte-Carlo Tree Search algorithm has been successfully applied in various domains. However, its performance heavily depends on the Monte-Carlo part. In this paper, we propose a generic way of improving the Monte-Carlo simulations by using RAVE values, which already strongly improved the tree part of the algorithm. We prove the generality and efficiency of our approach by showing improvemen...
متن کاملA Variable Neighbourhood Monte Carlo Search for Component Placement Sequencing of Multi-Head Placement Machine
This work considers the optimisation of component placement sequencing to improve the efficiency of theoretical multi-head surface mount device placement machines in printed circuit board assembly. We develop a Variable Neighbourhood Monte Carlo Search (VNMS), which employs a variable neighbourhood search technique with an Exponential Monte Carlo acceptance criterion. VNMS is a descent-ascent h...
متن کاملMonte Carlo Chess
MCC, a UCT based Chess engine, was created in order to test the performance of Monte-Carlo Tree Search for the game of Chess. Mainly by modifications that increase the accuracy of the simulation strategy, the performance of the base implementation was improved by approximately 864 Elo. MCC performed still too bad to compete with Minimax based chess programs or to seriously suffer from search tr...
متن کاملA Variable Neighbourhood Search for Component Pick-and- Place Sequencing in Printed Circuit Board Assembly
This work presents a heuristic for component pick-and-place sequencing to improve the throughput of a multi-head surface mount device placement machine for assembling printed circuit board. We present a Variable Neighbourhood Monte Carlo Search (VNMS), which employs variable neighbourhood search with an Exponential Monte Carlo acceptance criterion. VNMS is a descent-ascent heuristic that operat...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2006