[Paper Review] A Compact Self-organizing Cellular Automata-based Genetic Algorithm

Vasileios Barmpoutis, Gary F. Dargush|ArXiv.org|Nov 15, 2007

Cellular Automata and Applications45 references3 citations

TL;DR

This paper proposes a compact, self-organizing genetic algorithm (GA) that uses cellular automata (CA) principles to enable crossover only between neighboring individuals in a spatially structured population. By leveraging CA's self-organization rather than parallelism, the method achieves efficient search with minimal population size, reducing premature convergence and improving solution accuracy on test functions and truss design problems.

ABSTRACT

A Genetic Algorithm (GA) is proposed in which each member of the population can change schemata only with its neighbors according to a rule. The rule methodology and the neighborhood structure employ elements from the Cellular Automata (CA) strategies. Each member of the GA population is assigned to a cell and crossover takes place only between adjacent cells, according to the predefined rule. Although combinations of CA and GA approaches have appeared previously, here we rely on the inherent self-organizing features of CA, rather than on parallelism. This conceptual shift directs us toward the evolution of compact populations containing only a handful of members. We find that the resulting algorithm can search the design space more efficiently than traditional GA strategies due to its ability to exploit mutations within this compact self-organizing population. Consequently, premature convergence is avoided and the final results often are more accurate. In order to reinforce the superior mutation capability, a re-initialization strategy also is implemented. Ten test functions and two benchmark structural engineering truss design problems are examined in order to demonstrate the performance of the method.

Motivation & Objective

• To develop a compact genetic algorithm that maintains population diversity and avoids premature convergence using self-organizing principles.
• To integrate cellular automata (CA) rules into genetic algorithm operations, particularly crossover, to enable localized, neighbor-based evolution.
• To explore whether self-organization in CA can enhance mutation and exploration capabilities in small populations.
• To validate the approach on benchmark optimization problems, including structural engineering truss design.
• To improve solution accuracy and convergence speed compared to traditional GAs through re-initialization and compact population dynamics.

Proposed method

• Each individual in the GA population is mapped to a cell in a one-dimensional cellular automata (CA) lattice.
• Crossover operations are restricted to adjacent cells only, based on predefined CA transition rules.
• The CA rule set governs how genetic material is exchanged between neighboring individuals, promoting localized evolution.
• The algorithm leverages the inherent self-organizing behavior of CA to dynamically shape population structure and promote diversity.
• A re-initialization strategy is applied to revive stagnated individuals and enhance mutation-driven exploration.
• The population size is kept minimal—typically just a few individuals—focusing on compactness and efficiency.

Experimental results

Research questions

• RQ1 Can a compact GA with minimal population size achieve better convergence and accuracy than traditional GAs?
• RQ2 How does neighbor-based crossover governed by CA rules affect population diversity and search efficiency?
• RQ3 To what extent does the self-organizing nature of CA reduce premature convergence in genetic optimization?
• RQ4 Can re-initialization in a CA-structured population further enhance mutation and exploration?
• RQ5 How does the proposed method perform on complex, real-world engineering problems like truss design?

Key findings

• The proposed CA-based GA achieved superior convergence and solution accuracy on ten test functions compared to standard GA implementations.
• The compact population structure significantly reduced premature convergence, even with only a few individuals.
• The self-organizing CA rules enabled dynamic reconfiguration of the population, enhancing local search and diversity.
• The re-initialization strategy effectively revived stagnant individuals, increasing the algorithm's ability to escape local optima.
• On two benchmark truss design problems, the method produced more accurate and lightweight designs than conventional GAs.
• The algorithm demonstrated robust performance across diverse problem types, indicating strong generalization capability.