An Algebraic Approach to Population-Based Evolutionary Algorithm Generation

Evolutionary algorithms (EAs) are popular in solving a diversity of problems, but current algorithm design approaches typically require formulating an algorithmic structure for each individual problem. The paper presents an algebraic framework for high-level specification of general-purpose metaheuristic methods, which cover a wide range of population-based EAs. Based on specification composition and refinement, the framework support mechanical program generation for concrete problem solving. We illustrate the applications of the framework in two typical optimization problems, which show that the proposed approach can achieve a high level of abstraction and mechanization without losing performance.