Game Theory-Inspired Evolutionary Algorithm for Global Optimization

Many approaches that model specific intelligent behaviors perform excellently in solving complex optimization problems. Game theory is widely recognized as an important tool in many fields. This paper introduces a game theory-inspired evolutionary algorithm for global optimization (GameEA). A formul...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Algorithms Jg. 10; H. 4; S. 111
1. Verfasser: Yang, Guanci
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Basel MDPI AG 01.12.2017
Schlagworte:
adaptive learning
Chromosomes
Copying
Decision making
Evolutionary algorithms
game evolutionary algorithms (GameEA)
Game theory
Gene sequencing
genetic algorithm
Genetic algorithms
Global optimization
imitation
Learning
optimization problems
ISSN:1999-4893, 1999-4893
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Many approaches that model specific intelligent behaviors perform excellently in solving complex optimization problems. Game theory is widely recognized as an important tool in many fields. This paper introduces a game theory-inspired evolutionary algorithm for global optimization (GameEA). A formulation to estimate payoff expectations is provided, which is a mechanism to make a player become a rational decision-maker. GameEA has one population (i.e., set of players) and generates new offspring only through an imitation operator and a belief-learning operator. An imitation operator adopts learning strategies and actions from other players to improve its competitiveness and applies these strategies to future games where one player updates its chromosome by strategically copying segments of gene sequences from a competitor. Belief learning refers to models in which a player adjusts his/her strategies, behavior or chromosomes by analyzing the current history information to improve solution quality. Experimental results on various classes of problems show that GameEA outperforms the other four algorithms on stability, robustness, and accuracy.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1999-4893
1999-4893
DOI:10.3390/a10040111