Indian Journal of Science and Technology

Abstract

Background/Objectives: The article considers the special problem of distributing transport resources in real time for freight transportation companies that carry about loads according to the Full Truck Load (FTL) business model. Methods/ Statistical analysis: To solve the real-time factor problem multi-agent technologies have been used, which are successfully applied in many modern automation systems in transportation companies. During the research a large amount of statistical data obtained as a result of the operation of multi-agent automation systems were processed. Findings: The possibility to improve the quality of their resulting decisions was revealed. It was proposed to divide the problem solving process into 2 stages: the initial stage of resource allocation (static) and the real-time stage (dynamic). It was suggested to apply the Hungarian method at the static stage and the existing multi-agent method at the dynamic stage. To solve the problem at the static stage, it was brought to the classic assignment problem that can be solved by the Hungarian algorithm that has proven itself as a good accurate method. The article showed the impossibility of solving the problem by the same method at the dynamic stage, since even in the special case of the dynamic stage one cannot determine the correct resource for the next order. As a result, a new hybrid method for solving this problem has been developed. The results of effectiveness of this method were obtained using actual data on transportation companies as compared to the existing one, and significant efficiency gains were demonstrated. Applications/Improvements: The new method can be used in existing multi-agent automation systems in transportation companies to improve the quality of solutions.

Keywords: Assignment Problem, Full Truck Load (FTL), Hungarian Algorithm, Multi-agent Technology, Real Time