Algorithms for improved efficiency in transportation models

In this paper we develop algorithms for a variety of transportation problems which have not yet been widely studied. The first problem we study is that of scheduling signal lights at a traffic junction. The signal setting problem for a traffic junction involves setting the cycle time and the green times and sequencing them so as to maximize a measure of the flow thru the junction. The solution procedure we develop takes advantage of structural characteristics within the junction to find the best signal sequence and the optimal time for each green signal.; The second transportation problem we study is that of assigning dockside cargo crates to positions in a hold of a designated cargo ship. Cargo are initially stacked at a holding yard adjacent to the ship's berth. They must be loaded using a Gantry crane or transtainer system, one at a time, onto a flatbed truck which takes them to a position adjacent to the ship, where they are loaded into the hold of the ship. The objective of this problem is to find the order in which to transfer the crates from the yard stack to the ship. Our algorithm determines the loading schedule which optimizes port efficiency, while maintaining an acceptable level of ship stability. Our solution procedure involves solving a specially structured assignment problem repeatedly until we have a solution which satisfies stability requirements and maximizes port efficiency.; The third transportation problem we study is that of scheduling port cranes in an attempt to load or unload ships in the most efficient manner. The problem we study starts with several ships berthed at a port. Each ship has multiple holds which will require some work time to be loaded or unloaded. There are multiple shore cranes which must be shared among the ships and their holds. Our objective is to determine the crane work schedule which minimizes the total weighted delay times (and, therefore, costs) for the ships currently berthed. We develop an integer programming model for this problem.