Estimation of Distribution Algorithm for Optimization of System Reliability

As an important branch of evolutionary algorithms, an Estimation of Distribution Algorithm (EDA) uses some selected individuals to build a probability model to produce offspring by sampling the probability model, which is a process of statistically learning from the selected individuals and then probabilistically sampling the probability model. A redundancy optimization model is given and many optimization methods to solve optimum model and their advantages and shortages are analyzed. A new method based on estimation distribution algorithm for optimization of system reliability is put forward and better population selection proportions are analyzed. Compared with simulated annealing algorithm, genetic algorithm and particle swarm algorithm, the estimation distribution algorithm is more effective through result. This method has good scalability, and it can be modified solve general combinatorial optimization problems.