Intelligent function model discovery system based upon gene expression programming

Function Model Discovery System (FMDS) is an important component of Decision Support System (DSS). Traditional Intelligent Model Base system depends on transcendental knowledge, and it is difficult to implement vivid intelligentization. To solve the problem, this paper proposes Implicit Intelligent Model Mining Algorithms based on GEP and its revised algorithms (IIMM-GEP), and furthermore implements Intelligent Function Model Discovery System (IFMDS). The main contributions include: 1) The concepts of Explicit intelligent Model, Implicit Intelligent Model, and Explicit Gene are proposed. 2) The proposed algorithm IIMM-GEP, Explicit Intelligent Model Mining Algorithms Based on GEP (EIMM-GEP) and implements the Intelligent Function Model Discovery System (IFMDS) based on IIMM-GEP. 3) Studies the interface between the IFMDS and GIS, DSS. 4) It gives comparison experiment. The IFMDS is true intelligent Model Base constructing system base on it both the types and parameters of model are determined automatically by the IIMM-GEP. The comparison experiment results show that the EIMM-GEP method is compatible with the traditional methods, and moreover it is easy to submit the new type of the model while the traditional methods must write new program code to submit the new type of the model. The precision of the function model found by the IIMM-GEP is improved 14.9% than traditional method.