| Summary: | 碩士 === 國立臺灣科技大學 === 工業管理系 === 102 === A Load Plan determines the number of trailers required to move freight from one terminal to another in a time-space network, yet it does not deal with the problem of matching drivers to these trailers so as to execute such load plan and minimize the total mileage in the line haul network. In this thesis, a network used by a local Less-Than-Truckload motor carrier is considered to determine the best tour for individual drivers departing from their domicile and returning to it by the end of the shift. Drivers are assigned dispatches (movement of trailers between terminals) as required by the load plan, as well as resting periods at terminals and deadheading trips if needed.
The problem is modeled as a Driver Scheduling Problem embedded in a time-space hub-and-spoke type of network. Moreover, an Integer Programming-based local search heuristic named Clustering Method was designed to solve problems of different size using CPLEX as the integer programming solver. Furthermore, the solution quality and speed performance of this heuristic was compared with a rule-based approach, Period Sequence Method, which is simpler and computes very fast for small size instances. Experimental results demonstrate that the Clustering Method performs well on large-scale problems whereas Period Sequence Method can easily solve small instances within seconds.
|
|---|
