Exploring self-play as an alternative to coevolution
Journal: IPASJ International Journal of Computer Science (IIJCS) (Vol.5, No. 12)Publication Date: 2018-01-08
Authors : Yash Mundra;
Page : 6-10
Keywords : Keywords:coevolution; self-play; algorithm; chess;
Abstract
ABSTRACT This paper introduces a novel variant of self-play training which is used for optimizing functions whose size of parameter space computationally precludes the sustainedapplication ofconventional population coevolution for a large number of iterations. The proposed algorithm provides an effective way to fine-tune and adjust function parameters towards their optimal values. I use chess as a test bed but the underlying algorithm is applicable wherever self-competition can be effectively defined. The randomly initialized organisms undergo three phases of training (1) using conventional evolutionary algorithms (2) training using self-play (3) single population coevolution. The algorithm is intended to be used to provide similar performances as the SOTA but at a much less computational expenditure.
Other Latest Articles
- Smart Irrigation based on Cloud
- The Evaluation of Data sets with Computed Tomography Images using Image Processing Algorithms
- Differences of Environmental Conditions by the Type of ICT-based Horticulture Greenhouses
- LONG TERM CHARACTERIZATION OF FIBRE REINFORCED CONCRETE
- Review on Data Storage Telecommunication cloud based Model
Last modified: 2018-01-19 15:36:38