最短路径算法是计算机科学与地理信息科学领域的研究热点，而标号算法则是最短路径算法中的重要一族。长期以来，对于最短路径的算法实现，绝大多数都是围绕以Dijkstra算法为核心的标号设定算法来展开，而对标号改正算法的研究与应用却非常少见。为了对交通网络最短路径进行更有效、更快速的计算，通过对标号改正算法思想的深入分析，针对其中最具代表性的Pallottino算法，从存储结构和运行结构两方面进行了算法的优化改进，同时分析了该算法的时间复杂度和空间复杂度，并利用实际的大规模城市交通网络进行了效率测试。结果显示，与目前公认最优的标号设定算法中基于逼近桶结构的Dijkstra算法相比，该改进的标号改正Pallottino算法具有更好的适用性和更高的运行效率，因此在交通网络最短路径分析应用中具有很高的应用价值。

The shortest path problem, in which label algorithms are of consequence, is a research topic in the field of geographic information science and computer science. Among label algorithms, label setting algorithms in which Dijkstra occupies the core position are always considered as the first choice in applications while label correcting algorithms are rarely used. In this paper, followed by the theory expatiation on label correcting algorithms, an optimization of the wellknown Pallottino algorithm is set forward on the basis of both storage and implementation structures. The author then analyzes its time complexity and spatial complexity and finally tests its actual efficiency in Beijing transportation network, It proves a better implemental efficiency and applicability comparing with the best label setting algorithm- Dijkstra algorithm implemented with approximate buckets, As a result, it provides a good choice for network related analysis works and should be particularly useful to researchers and practitioners in operations research, transportation, and GIS,