A simulation optimization method for upper and lower bound convergence search with screening mechanism

Abstract

The invention relates to an upper and lower limit convergence search simulation optimization calculation method with a screening mechanism. To maximize the overall service satisfaction of the service system as the optimization goal, establish a mixed integer optimization model of the service system with two heterogeneous service types with a screening mechanism, and then use the upper and lower bounds to continuously search for the system simulation value in a systematic way , change the upper or lower bound of the search in each iterative calculation, and then find a new threshold value through the dichotomy until it converges to the waiting time value that meets the confidence interval. The method of the invention can find out the approximate best solution of the optimization model under the limitation of large solution space, time and sampling cost.

Description

technical field [0001] The invention relates to a simulation optimization calculation method for upper and lower bound convergence search with a screening mechanism. Background technique [0002] The simulation optimization methods of the prior art focus on developing the relevant theory of how to converge to the optimal solution when the number of samples approaches infinity. Feasible solutions of stochastic constraints. [0003] In practice, however, the size of the sample usually depends on the subjective decision of the decision maker, and only a limited number of samples can be used. At the same time, the larger the number of samples required for system simulation, the greater the sampling cost; and in the case of insufficient samples, simulation optimization cannot guarantee that a feasible solution can be found. In addition, in dealing with optimization problems with discrete variables, it is necessary to estimate the subgradient function (Subgradient). The research...

AI Technical Summary

Problems solved by technology

[0002] The simulation optimization methods in the prior art focus on the development of related theories on how to converge to the optimal solution when the number of samples approaches infinity. For the simulation optimization model, in the limited number of samples, it is impossible to guarantee that all satisfying Feasible solution of stochastic constraints

[0003] However, in practice, the size of the sample usually depends on the subjective decision of the decision maker, and only a limited number of samples can be used

At the same time, when the number of samples required for system simulation is larger, it means that the sampling cost is higher; and in the case of insufficient number of samples, simulation optimization cannot guarantee that a feasible solution can be found.

In addition, in dealing with optimization problems with discrete variables, it is necessary to estimate the subgradient function (Subgradient). The research work in the past literature often uses the finite difference method (FiniteDifferences), which often requires a huge calculation and simulation time cost.

Images