Multi-Index Dynamic Optimization Method for Enterprise Evaluation Based on Correlation Constraints

Abstract

Most of the current enterprise evaluation methods are based on a comprehensive index system. However, there are significant differences in the evaluation effectiveness of indicators and the difficulty of obtaining index data, which determines that a unified and comprehensive Index system cannot be adopted for different enterprises. Therefore, how to dynamically screen out the index with the least acquisition cost, meet the evaluation needs and not destroy the index relationship from the index system of related fields has become a significant problem. The current research results usually assume that the index is independent of each other and rarely consider the correlation constraints between different indexes. For this reason, this paper proposes a multi-index optimization problem based on correlation constraints (Multi-Index Optimization Based on Relevance Constraints, MIO-RC). MIO-RC is based on the evaluation value of each index, and the cost of index acquisition then selects the index set with the minimum cost that satisfies the expected contribution value. MIO-RC problem defines the correlation constraints between various evaluation indicators and defines a dynamic contribution value function based on enterprise characteristics and correlation indicators. In this paper, the MIO-RC problem is an NP-hard problem, and an IR-GA algorithm based on a genetic algorithm is designed to solve the MIO-RC problem under different constraints. Finally, this paper conducts a verification analysis based on many enterprise data sets. The experimental results show that: in terms of performance, compared with the simulated annealing algorithm and the particle swarm evolution algorithm, the IR-GA algorithm converges faster.