Algorithmic Mechanism Design of Evolutionary Computation

Abstract EN

We consider algorithmic design, enhancement, andimprovement of evolutionary computation as a mechanism designproblem.

All individuals or several groups of individualscan be considered as self-interested agents.

The individuals inevolutionary computation can manipulate parameter settings andoperations by satisfying their own preferences, which are definedby an evolutionary computation algorithm designer, rather thanby following a fixed algorithm rule.

Evolutionary computationalgorithm designers or self-adaptive methods should constructproper rules and mechanisms for all agents (individuals) toconduct their evolution behaviour correctly in order to definitelyachieve the desired and preset objective(s).

As a case study,we propose a formal framework on parameter setting, strategyselection, and algorithmic design of evolutionary computation byconsidering the Nash strategy equilibrium of a mechanism designin the search process.

The evaluation results present the efficiency of the framework.

This primary principle can be implemented inany evolutionary computation algorithm that needs to considerstrategy selection issues in its optimization process.

The finalobjective of our work is to solve evolutionary computation designas an algorithmic mechanism design problem and establish itsfundamental aspect by taking this perspective.

This paper is thefirst step towards achieving this objective by implementing astrategy equilibrium solution (such as Nash equilibrium) inevolutionary computation algorithm.