In this paper,the nonlinear optimization problems with inequality constraints are discussed. Combining the ideas of the strongly sub-feasible directions method and theε-generalized projection technique,a new algorith...In this paper,the nonlinear optimization problems with inequality constraints are discussed. Combining the ideas of the strongly sub-feasible directions method and theε-generalized projection technique,a new algorithm starting with an arbitrary initial iteration point for the discussed problems is presented.At each iteration,the search direction is generated by a new ?-generalized projection explicit formula,and the step length is yielded by a new Armijo line search.Under some necessary assumptions, not only the algorithm possesses global and strong convergence,but also the iterative points always get into the feasible set after finite iterations.Finally,some preliminary numerical results are reported.展开更多
In this paper, a new superlinearly convergent algorithm for nonlinearly constrained optimization problems is presented. The search directions are directly computed by a few formulas, and neither quadratic programming ...In this paper, a new superlinearly convergent algorithm for nonlinearly constrained optimization problems is presented. The search directions are directly computed by a few formulas, and neither quadratic programming nor linear equation need to be sovled. Under mild assumptions, the new algorithm is shown to possess global and superlinear convergence.展开更多
面向能耗优化的面积(核数)-功率(频率)分配问题是当前众核处理器研究热点之一.通过性能-功耗模型了解其在核数-频率空间的分布规律,然后在核数和频率级别这2个维度上通过实测执行逐步搜索,可以获取"核数-频率"配置的最优解,...面向能耗优化的面积(核数)-功率(频率)分配问题是当前众核处理器研究热点之一.通过性能-功耗模型了解其在核数-频率空间的分布规律,然后在核数和频率级别这2个维度上通过实测执行逐步搜索,可以获取"核数-频率"配置的最优解,从而达到能耗优化的目的;然而本领域现有方法在核数-频率空间内实测搜索最低能耗时收敛速度慢、搜索开销大、可扩展性差.针对此问题,提出了一种基于求解最优化问题的经典数学方法——可行方向法的最低能耗搜索方法(energy-efficient optimization based on feasible direction method,EOFDM),每次执行都能从核数和频率2个维度上同时减小搜索空间,在迭代执行中快速收敛至最低能耗点.该方法与现有研究中最优的启发式爬山法(hill-climbing heuristic,HCH)进行了对比实验,平均执行次数、执行时间和能耗分别降低39.5%,46.8%,48.3%,提高了收敛速度,降低了搜索开销;当核数增加一倍时,平均执行次数、执行时间和能耗分别降低48.8%,51.6%,50.9%;当频率级数增加一倍时,平均执行次数、执行时间和能耗分别降低45.5%,49.8%,54.4%,在收敛速度、搜索开销和可扩展性方面均有提高.展开更多
指出准协同优化策略(Collaborative optimization,CO)存在的数值缺陷及其原因。针对系统级优化不满足Kuhn-Tucker条件所导致的计算困难,提出一种基于可行方向序列无约束极小化技术(Feasible direction sequential unconstrainedminimiza...指出准协同优化策略(Collaborative optimization,CO)存在的数值缺陷及其原因。针对系统级优化不满足Kuhn-Tucker条件所导致的计算困难,提出一种基于可行方向序列无约束极小化技术(Feasible direction sequential unconstrainedminimization technology,FD-SUMT)外点法的改进协同优化策略(Enhanced collaborative optimization with FD-SUMT method,ECO-FSM)。在系统级优化中使用FD-SUMT外点法,该方法不依赖Lagrange乘子并且能够将系统级设计变量限定在设计变量可行域内,避免传统SUMT外点法设计变量越界所导致的异常。利用学科间动态不一致信息更新系统级优化中的罚因子以加速学科间的协调。利用测试问题检验ECO-FSM的性能,并与其他的CO进行比较研究。研究结果表明ECO-FSM消除了系统级优化中设计变量越界的现象,收敛性、数值稳定性以及收敛速度得以显著提高。将ECO-FSM用于亚声速喷气式客机总体方案优化设计,优化结果表明ECO-FSM具有工程实用性。展开更多
基金supported by the National Natural Science Foundation of China under Grant Nos.71061002 and 10771040the Project supported by Guangxi Science Foundation under Grant No.0832052Science Foundation of Guangxi Education Department under Grant No.200911MS202
文摘In this paper,the nonlinear optimization problems with inequality constraints are discussed. Combining the ideas of the strongly sub-feasible directions method and theε-generalized projection technique,a new algorithm starting with an arbitrary initial iteration point for the discussed problems is presented.At each iteration,the search direction is generated by a new ?-generalized projection explicit formula,and the step length is yielded by a new Armijo line search.Under some necessary assumptions, not only the algorithm possesses global and strong convergence,but also the iterative points always get into the feasible set after finite iterations.Finally,some preliminary numerical results are reported.
文摘In this paper, a new superlinearly convergent algorithm for nonlinearly constrained optimization problems is presented. The search directions are directly computed by a few formulas, and neither quadratic programming nor linear equation need to be sovled. Under mild assumptions, the new algorithm is shown to possess global and superlinear convergence.
文摘面向能耗优化的面积(核数)-功率(频率)分配问题是当前众核处理器研究热点之一.通过性能-功耗模型了解其在核数-频率空间的分布规律,然后在核数和频率级别这2个维度上通过实测执行逐步搜索,可以获取"核数-频率"配置的最优解,从而达到能耗优化的目的;然而本领域现有方法在核数-频率空间内实测搜索最低能耗时收敛速度慢、搜索开销大、可扩展性差.针对此问题,提出了一种基于求解最优化问题的经典数学方法——可行方向法的最低能耗搜索方法(energy-efficient optimization based on feasible direction method,EOFDM),每次执行都能从核数和频率2个维度上同时减小搜索空间,在迭代执行中快速收敛至最低能耗点.该方法与现有研究中最优的启发式爬山法(hill-climbing heuristic,HCH)进行了对比实验,平均执行次数、执行时间和能耗分别降低39.5%,46.8%,48.3%,提高了收敛速度,降低了搜索开销;当核数增加一倍时,平均执行次数、执行时间和能耗分别降低48.8%,51.6%,50.9%;当频率级数增加一倍时,平均执行次数、执行时间和能耗分别降低45.5%,49.8%,54.4%,在收敛速度、搜索开销和可扩展性方面均有提高.
