Optimization with Conflicting Objectives in Logistic Services

We study a single period supply chain model with deterministic demand at the destination and multiple paths through which supplies can be sent. Each path may fail independently. When a path fails, the supply that is sent through that path does not arrive at the destination. We wish to minimize the expected shortage, but also minimize the total supply that is used. We investigate the behavior of various heuristics, deriving some general rules for making good choices and proving an advantage to complex optimization techniques over simple heuristics. We show that when multiple paths have the same characteristics, the optimal solution can be computed significantly faster.