We consider the network resource allocation problem, faced by sea freight operators, for standardized cargo transportation networks with time-dependent and stochastic demand. Standardized cargo capacity is sold through allotment contracts and on the spot market. The objective is to decide on the allotment contracts to select at the beginning of the booking horizon, as well as investigate dynamic booking control policies based on which spot market requests are accepted (booked) over the booking horizon before departure. This problem is formulated and solved in two stages: In the first stage, the allotment selection problem is formulated as an integer program in which the objective is to maximize the expected profit generated from the realized utilization of each resource in the network. The optimal solution to the allotment selection problem becomes the input to the spot booking control problem in the second stage. We formulate the spot booking problem as a dynamic program, and provide an efficient heuristic as an alternative to the computationally prohibitive dynamic programming (DP) approach. The DP is solved optimally for a small-size problem and for a variety of numerical cases. For large size problems, heuristics are proposed and tested via simulation. Also extensive sensitivity analysis is performed to test the heuristics, gain managerial insights and provide practical recommendations. We recommend a well-performing heuristic that outperforms the trivial first-come first-serve policy for large-sized problems. This leading heuristic is within 2.75%, on average, of the optimal DP solution for small problems and 3.9% for large problems.

我们考虑了海运运营商面临的网络资源分配问题，用于具有时间依赖和随机需求的标准化货物运输网络。标准化货运能力通过配股合同和现货市场出售。目标是决定在预订期限开始时选择的分配合同，并根据出发前在预订期限内接受（预订）哪些现货市场请求来调查动态预订控制策略。这个问题分两个阶段制定和解决：在第一阶段，分配选择问题被制定为一个整数程序，其目标是最大化网络中每个资源的实现利用所产生的预期利润。分配选择问题的最优解成为第二阶段现货预订控制问题的输入。我们将定点预订问题制定为一个动态程序，并提供一种有效的启发式方法，作为计算上禁止的动态规划 (DP) 方法的替代方法。DP 是针对小规模问题和各种数值情况的最佳解决方案。对于大型问题，通过模拟提出并测试启发式方法。还进行了广泛的敏感性分析以测试启发式方法，获得管理见解并提供实用建议。我们推荐一种性能良好的启发式算法，它优于针对大型问题的微不足道的先到先服务策略。这种领先的启发式算法平均在小问题的最佳 DP 解决方案的 2.75% 以内，3。