A Fireworks-inspired Estimation of Distribution Algorithm
Hao Li · IOP Conference Series Materials Science and Engineering · 2019
Abstract Estimation of distribution algorithm(EDA) is a popular evolutionary algorithm which is obtained widely attention. On the basis of EDA, combining with the idea of fireworks algorithm, we proposed a fireworks-inspired multi-model EDA(FMEDA). In FMEDA, multiple probabilistic models are used and generate different number of solutions by introducing the idea of firework explosion. Next, each model evolves by compared with its own best solution directly. The method of setting the limit of probability is taken to prevent from premature convergence. Applying this algorithm to the function optimization problem and the knapsack problem, the experimental result shows that the proposed algorithm has better performance than CGA and PBIL.