Usually, since is a random vector w

Usually, since is a random vector with unknown distribution, it is impossible to convert the above P-model into an analytic model, and thus stochastic stimulation is an alternative tool to solve one such uncertain programming problem. Additionally, based on the law of large number, when the sample sizes, attached by candidates in are large enough, an empirically optimal reliable solution will sufficiently approach the true optimal reliable solution. If so, it is inevitable to cause computationally high complexity. Conversely, if each candidate solution is attached a small sample size, it is easy to deem inferior candidates as superior ones in the process of solution search, and as a result the optimized quality is influenced seriously by noise interference strength. Therefore, in order to efficiently acquire an approximate optimal reliable solution to the above problem, in this paper we require that the sample size of at each candidate solution x, n(x), be determined dynamically, by which the probability estimates of the chance constraints can be estimated. Under sample sizes m and Mn, the objective value of x, , can be estimated by Algorithm 1 above, where Mn depends on the sampling moment of n.