On solving the cross-dock door assignment problem
针对交叉转运站中货物运输成本最小化的门分配问题，提出一种线性化整数规划重构模型，通过分解子模型和线性搜索启发式算法，在55个实例上验证了该方法在求解质量和时间上优于CPLEX，尤其适用于大规模实例。
A class of strong lower bounds on the solution value of a Linearised Integer Programming reformulation is introduced for the binary quadratic optimisation model to assign origin and destination nodes to strip and stack doors, resp., in a cross-dock infrastructure. The goal is to minimise the transportation cost of the commodities to be handled at the cross-dock. Strip- and stack-related decomposition submodels are developed by taking benefit of the Integer Linearisation Property that appears in the new model. A linear search heuristic is also provided for obtaining feasible solutions by exploiting the special structure of the problem. We present an extensive computational study on a testbed of 55 instances to show that the proposed joint scheme for lower bounding and feasible solution providing is very efficient. The results obtained by our proposal are similar to those obtained by CPLEX in the 50 instances taken from the literature, but usually the time is much smaller for the large instances. The comparison with CPLEX is particularly good for the new five largest instances, which make it very promising for its application in real-life cases.