Research On The Method Of Trust Evaluation And Service Combination Based On Trust In Supply Chain Enterprises

With the development of science and technology level and changing market demand, the supply chain cooperation between enterprises, more and more members in order to enable enterprises can benefit, and let every business has a strong competitive edge, it must Composition and operation of the supply chain in-depth study, in this context, the study of supply chain and supply chain management is becoming a hot spot. The current corporate partnerships issues of more and more attention, but mainly to emphasize technical elements and processes, while ignoring a vital role to play in the long-term factors in- trust. In the prior study, the factors affecting corporate trust, trust in research evaluation methods and choose credible enterprise services, trusted enterprise service portfolio and other issues, there are still shortcomings among enterprises, for the above problem, this paper research work as follows.Firstly, a comprehensive trust evaluation method is proposed to improve the accuracy of the trading trust assessment. The method takes into consideration four indicators, namely degree of transaction satisfaction, transaction value, number of transaction and transaction time. It also incorporates a punishment mechanism to curb fraudulent transactions by enterprises. In addition, the model brings in a recommendation trust mechanism and proposes a trust evaluation method based on recommendation reliability. The above two trust values are combined to reach a comprehensive enterprise trust value, the validity of which will be subjected to simulations.Secondly, to improve the comprehensiveness and feasibility of enterprise trust evaluation, a comprehensive trust evaluation method is proposed that can fully assess enterprise trust. This method firstly assesses the basic enterprise trust that basic properties of the enterprise demonstrate, and then construct an indicator system and an assessment method based on BP neural network, thus quantizing basic enterprise trust. A comprehensive enterprise trust value is obtained in combination with the comprehensive enterprise transaction trust and basic enterprise trust. Simulated experiments are conducted to verify the feasibility of the method.Thirdly, a trust-based service selection method is proposed for service selection in view of reduced degree of service satisfaction resulting from fraudulent services. In this method, available services are screened by their service trust and a fuzzy clustering is proposed in consideration of four properties, namely service quality, price, time and trust. To meet clients’ different service preferences, weight coefficients are used to set service preferences. Services of high service evaluation value will be selected to obtain higher customer satisfaction with the service. Simulated experiments show that the customized service selection method based on trust results in improved service satisfaction compared with both service selection based on trust and random service selection.Finally, due to the fact that malicious services, if involved in the service composition, will reduce the actual service quality of the whole portfolio, a trust-based service composition method is proposed. The method makes some improvement to the traditional genetic algorithm, optimizes the original one-dimensional integer coding scheme, and proposes a setting method for adaptive crossover and mutation probabilities. The method also introduces a dynamic fitness function, based on which a self-corrective dynamic fitness function is proposed. Corrections can make the service quality, service price and service time closer to the actual level. Simulation results show that the proposed method improves the time efficiency of the service composition, the global searching ability and the actual fitness value in different degrees with respect to the existing methods.