中文摘要：

引入服务水平等级概念,以OD流量最大和运输费用最小为目标,构建铁路货物运输网络能力计算多目标规划模型。采用按服务水平等级由高至低顺序进行车流量调整的策略,设计基于K短路和最小费用流问题的启发式算法;并针对最小费用流问题,分别给出基于Lingo软件和遗传算法的2种求解方法。在遗传算法中,对染色体采用二进制编码和运用Double-Sweep算法计算各支车流在给定服务水平等级下的可选径路,利用交叉、变异操作遍历可能的车流路径集合,使用启发式车流调整策略,实现线路及车站能力限制条件下不同路径集合的车流分配。算例测试表明,基于Lingo软件的算法适合于求解小规模问题,而基于K短路的遗传算法更适用于求解大规模问题,且具有较高的稳定性和适应性。

英文摘要：

Based on the concept of service level, with the maximum OD flow and the minimum transportation cost as the target, a multi-objective model for calculating the transportation capacity of railway freight transportation network is constructed. By adopting the car flow adjusting strategy, in which the service level is in descending order, a heuristic algorithm based on K-shortest path and the minimum cost flow sub-problem is designed. Aiming at solve the minimum cost flow problem, two solution methods are pres- ented based on Lingo 8. 0 and genetic algorithm （GA） respectively. In GA, binary encoding and Double- Sweep algorithm are adopted for the chromosome to calculate the possible routes of all OD flows under the given service level. With crossover and mutation operations, all potential route sets might he generated and searched. The heuristic car flow adjustment strategy is employed to realize the car flow distribution of different path set under the limiting conditions of line and station capacity. The paper also gives several numerical examples to test the heuristic method. The testing results show that the heuristic method has strong adaptability and stability. Lingo can be used to solve small-size problems while genetic algorithm with K-shortest path is appropriate for large-size problems. Thus, a flexible and practical method for solving large-size railway freight network capacity problem is provided.