Particle swarm optimization based Nonlinear Least-squares Parameter Estimation of Maintenance Time Distribution

Maintenance time is the basis of maintainability quantitative analysis, and it is widely used in maintainability verification, demonstration or evaluation. The nonlinear least squares parameter evaluation methods of maintenance time distribution based on particle swarm optimization is proposed. Parameter's confidence interval is chosen as coding space of each particle, and objective function of nonlinear least-squares estimation is taken as fitness function of particle swarm optimization. With the operation of update of particle's velocity and position iteration iteratively, the optimum value of parameter can be acquired. Typical maintenance time distribution is taken Application instances, and the result shows that estimating precision of method proposed here is better than traditional method such as maximum likelihood estimation in most situations.