An important issue for providing better guarantees of Quality of Service (QoS) to applications is QoS rout-ing. The task of QoS routing is to determine a feasible path that satisfies a set of constraints while maintai...An important issue for providing better guarantees of Quality of Service (QoS) to applications is QoS rout-ing. The task of QoS routing is to determine a feasible path that satisfies a set of constraints while maintaining high u-tilization of network resources. For the purpose of achieving the latter objective additional optimality requirementsneed to be imposed. In general, multi-constrained path selection problem is NP-hard so it cannot be exactly solved inpolynomial time. Accordingly heuristics and approximation algorithms with polynomial or pseudo-polynomial timecomplexity are often used to deal with this problem. However, many of these algorithms suffer from either excessivecomputational complexity that cannot be used for online network operation or low performance. Moreover, they gen-erally deal with special cases of the problem (e. g. , two constraints without optimization, one constraint with opti-mization, etc. ). In this paper, the authors propose a new efficient algorithm (EAMCOP) for the problem. Makinguse of efficient pruning policy, the algorithm reduces greatly the size of search space and improves the computationalperformance. Although the proposed algorithm has exponential time complexity in the worst case, it can get verygood performance in real networks. The reason is that when the scale of network increases, EAMCOP controls effi-ciently the size of search space by constraint conditions and prior queue that improves computational efficiency. Theresults of simulation show that the algorithm has good performance and can solve effectively multi-constrained opti-mal path (MCOP) problem.展开更多
文摘An important issue for providing better guarantees of Quality of Service (QoS) to applications is QoS rout-ing. The task of QoS routing is to determine a feasible path that satisfies a set of constraints while maintaining high u-tilization of network resources. For the purpose of achieving the latter objective additional optimality requirementsneed to be imposed. In general, multi-constrained path selection problem is NP-hard so it cannot be exactly solved inpolynomial time. Accordingly heuristics and approximation algorithms with polynomial or pseudo-polynomial timecomplexity are often used to deal with this problem. However, many of these algorithms suffer from either excessivecomputational complexity that cannot be used for online network operation or low performance. Moreover, they gen-erally deal with special cases of the problem (e. g. , two constraints without optimization, one constraint with opti-mization, etc. ). In this paper, the authors propose a new efficient algorithm (EAMCOP) for the problem. Makinguse of efficient pruning policy, the algorithm reduces greatly the size of search space and improves the computationalperformance. Although the proposed algorithm has exponential time complexity in the worst case, it can get verygood performance in real networks. The reason is that when the scale of network increases, EAMCOP controls effi-ciently the size of search space by constraint conditions and prior queue that improves computational efficiency. Theresults of simulation show that the algorithm has good performance and can solve effectively multi-constrained opti-mal path (MCOP) problem.
文摘针对蝴蝶优化算法(butterfly optimization algorithm,BOA)在复杂环境路径规划过程中求解最短路径时存在收敛速度慢、易陷入局部最优等缺点,提出一种改进的蝴蝶优化算法。首先,在初始化蝴蝶种群时,为保证初代种群多样化,避免陷入局部最优解,通过Tent映射生成初代种群位置;其次,在蝴蝶香味计算阶段引入动态感觉模态,随着迭代过程的持续推进逐步增强蝴蝶的香味值,以缩短收敛时间;再次,为进一步缩短收敛时间,在全局搜索阶段引入遗传算法中的选择因子加快蝴蝶在全局搜索时向最优蝴蝶移动的速度;然后,在局部搜索阶段引入动态变异因子,有效避免在路径规划时陷入局部最优;最后,使用一种基于视线(line of sight,LOS)检测方法的初始种群生成策略,以进一步减少路径中断点的生成,同时确保由BOA算法生成的路径可行解的多样性。实验结果表明,改进的蝴蝶优化算法具有较快的收敛速度,且规划出来的路径在保证路径长度合理的情况下具有更高的平滑度。
