Adaptation in natural and artificial systems
Adaptation in natural and artificial systems
Sizing Populations for Serial and Parallel Genetic Algorithms
Proceedings of the 3rd International Conference on Genetic Algorithms
Introduction to Evolutionary Computing
Introduction to Evolutionary Computing
The gambler's ruin problem, genetic algorithms, and the sizing of populations
Evolutionary Computation
No free lunch theorems for optimization
IEEE Transactions on Evolutionary Computation
AICI '09 Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence
AICI '09 Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence
International Journal of Bio-Inspired Computation
Chaotic populations in genetic algorithms
Applied Soft Computing
Stochastic Populations, Power Law and Fitness Aggregation in Genetic Algorithms
Fundamenta Informaticae
Hi-index | 0.00 |
Biological populations are dynamic in both space and time, that is, the population size of a species fluctuates across their habitats over time. There are rarely any static or fixed size populations in nature. In evolutionary computation (EC), population size is one of the most important parameters and it received attention from EC pioneers from the very beginning. Despite many attempts to optimize the population sizing, the prevailing scheme in EC is still possibly the simplest --- the fixed size population. This is in strong contrast with population entities in nature. In this paper, we explore the effects of dynamic (fluctuating) populations on the performance of genetic algorithms (GA). In particular, we test five dynamic population-sizing patterns: random fluctuating population, increasing population, decreasing population, bell-shaped population, and inverse bell-shaped population and compare them against the fixed size population. Our experiment shows very promising results that the dynamic populations perform more efficiently than the traditional fixed size populations, in terms of the number of fitness function evaluations and memory space requirements. We also analyze why the dynamic populations should perform superior to the fixed size populations from the biological perspective.