ELM-Based Nonlinear Combination Intelligent Evaluation Approach For Candidate Partners’ Achievement Of Dynamic Alliance

Dynamic Alliance is a new kind of enterprise organization mode, and it’s a main mode of production operation and market competition at the age of information. For enterprise Dynamic Alliances, choosing a suitable partner is the key factor to success. Therefore, there is a strong theoretical and practical significance to research dynamic alliance partner selection strategies and its methods. As there are nonlinear relationship between the factors of affecting the performance of partners and decision goal, the traditional evaluation methods are often difficult to obtain the ideal effect. In recent years, some soft computing methods have been introduced to the performance evaluation of candidate partners modeling. In order to make full use of the information provided by the individual models, improve forecasting accuracy and enhance forecasting stability, this paper attempts to integrate a variety of soft computing methods, to establish a non-linear combination model based on extreme learning machine for the dynamic alliance candidate partners performance intelligently evaluated. The main work is as follows:(1)Taking to build a dynamic alliance partners candidate performance evaluation for example, it briefly introduces the main factors of affect the performance of dynamic alliance partners and common evaluation methods. In order to eliminate redundant information among the input data, speed up computing speed and improve the efficiency of the performance evaluation, we extract the original data characteristics by the principal component analysis, and eliminate information overlap between indexes.(2) On the basis of the candidate projects dynamic alliance partner performance evaluation system and data preprocessing on, we establish the four types of construction project dynamic alliance candidate partner performance evaluation model by using multiple linear regression, the GM(1, N), RBF neural network and SVM method respectively, and carry out computer simulation experiments.(3) Based on four kinds of single model, we attempt to establish a model of nonlinear combination based on ELM to evaluate intelligently the performance of dynamic alliance candidate partners, and compare it with the evaluation results of multiple linear regression model, GM(1, N) model, RBF neural network model and SVM model, the experimental results show that the effectiveness of the ELM nonlinear intelligent combination evaluation model.