The Generalized Expectation Value Model of Stochastic Programming Problem

In this paper, by analyzing the essential characteristic of stochastic programming and the deficiencies of the existing methods, we propose the concept of synthesizing effect function for the comparison of the objective function value, give an axiomatic system for stochastic synthesizing effect function, establish objective evaluation principles based on synthesizing effect; for the satisfaction of constraints, we propose a quasi-linear model based on expectation and variance; further we establish an operable stochastic programming model (called the generalized expected value model, denoted by GEVM for short), and analyze the performance through an example. The results indicate that our method not only includes the existing methods for stochastic programming, but also effectively merges the decision preferences into the solution, and extends and enriches the existing stochastic programming theory, so it can be widely used in many fields such as complicated system optimization and artificial intelligence.