Weight Calculation of GSM-R Network Quality of Service Evaluation Indexes Based on Analytic Hierarchy Process and Entropy Weight Method

Abstract

In order to assign an objective and compre-hensive weight to each index for evaluating the GSM-R network quality of service, this paper used the combination of subjective and objective weight value to overcome the singularity of determining the weight from one aspect. The entropy weight method was used to determine the objective weight, which was assigned to each index by the entropy calculation. The subjective weight was calculated on the basis of the improved analytic hierarchy process (AHP) and the scale construction method was used to construct the judgment matrix, which reduced the computational complexity and redundancy. In addition, three kinds of calculation methods were proposed for combining multiple experts’ opinions. Combining the subjective and objective weight vectors obtained by different combination methods, the results of multiple integrated weight vectors were obtained. Then the average value was calculated for each index. By comparing the correlation coefficient between each integrated weight vector and the average weight vector, a better multiple combining calculation method was determined to calculate the weight of each index in the evaluation system. This paper combined the actual test data of ZX section of XL high-speed railway and the experts’ questionnaires to obtain the weight assignment of GSM-R network quality of service indexes. It can be seen from the results that the voice quality, switching success ratio and average handover delay account for a large proportion, which is in line with reality. The judgment method of weight proposed in this paper is objective and feasible.