| Popis: |
In this paper there is introduced a comparative analysis of multi-layer data model used for parallelization of transportation network objects processing in the process of logistics optimization. A concept of multi-layer parallelization is based on distribution of the objects and subjects of a complex transportation logistics system including the road network, pick-up and delivery points, drivers, tracks and trailers, transportation instructions and order flows to the logical layers established considering the requirements of parallel computing. This paper continues this research by study of parallel computing characteristics comparing to traditional approach. The proposed data model allows increasing the performance of parallel computing application in transportation logistics problem domain. The considered performance metrics include execution time, total parallel overhead, speedup, and efficiency. Research results are implemented and applied in practice as a part of a transportation management software solution. |
