Military logistics planning is a complex and time consuming process, and ultimately many factors can impact the devised plan. We have developed a way of assessing the robustness of a logistics plan and associated supply chain by considering the occurrence of disruptive negative events. These events may occur in isolation or simultaneously. Each event is associated with a plan variable and the impact of a negative event can then be assessed by measuring whether the modification of its associated variable causes any stockouts in the plan during simulation. By modifying any two variables simultaneously, we can identify the minimum values of those two variables, which in combination cause a stockout. These minimum values are called breakpoints, and a set of breakpoints is referred to as a breakpoint front. We have investigated various approaches of calculating or estimating breakpoint fronts, including brute force search, the Monte Carlo method, 2D binary search and the multi-objective optimization methods, SPEA2 and NSGA-II. Our experiments showed that 2D binary search performed best overall, due to its low computation time and accuracy in calculating a range of breakpoint fronts.
展开▼
