利用遗传算法研究了动态并行路径流量工程的实现方法和可实现性.首先给出并行路径流量工程理论最优解的求解方法,并对其运算成本和运算结果在工程上的不可实现性进行了分析和验证;然后采用简洁、典型的分流方案,对已有的理论最优解进行局部近似,并基于这些分流方案从全局的角度设计遗传算法,达到了与理论最优解近似的优化效果,且运算量大为减小.通过仿真分析得出了在优化过程中,需要对并行路径间的分流误差进行有效控制的结论.
The implementation and practicability of dynamic parallel paths traffic engineering (DPPTE) with genetic algorithm was analyzed. An algorithm for the theoretical optimization of DPPTE was offered at first. The computing cost and the outcome were unrealized technically. Thereafter, a new genetic algorithm was designed based on practicable traffic assignment solutions, and the optimization performance of the algorithm was closed to the theoretical one with a little computing cost. The simulations proves that traffic assignment errors between parallel paths will be controlled in the optimizing.