考虑灵敏度区域的多目标鲁棒性优化算法

随机变量导致工程问题具有不确定性.设计者希望设计方案不仅能满足目标性能最优,而且希望目标性能受不确定性的影响在可接受范围之内.对此,本文提出了考虑灵敏度区域的多目标鲁棒性优化方法(multi-objective robust optimization based on sensitivity region,SR–MORO).SR–MORO可以用来解决设计变量存在不确定性时目标鲁棒性优化设计问题.该方法假定不确定性变量属于区间变量,并不需要知道随机变量的概率分布.SR–MORO采用非梯度优化方法,所以,它可以解决目标函数和约束条件不连续的情况.当参数变化幅度大,超过目标函数线性变化范围,该方法也同样适用.最后,通过实例验证了本方法的适用性.

Robust H_∞ directional control for a sampled-data autonomous airship

A robust H∞ directional controller for a sampled-data autonomous airship with polytopic parameter uncertainties was proposed. By input delay approach, the linearized airship model was transformed into a continuous-time system with time-varying delay. Sufficient conditions were then established based on the constructed Lyapunov-Krasovskii functional, which guarantee that the system is mean-square exponentially stable with H∞ performance. The desired controller can be obtained by solving the obtained conditions. Simulation results show that guaranteed minimum H∞ performance γ=1.4037 and fast response of attitude for sampled-data autonomous airship are achieved in spite of the existence of parameter uncertainties.

城市道路交通网络脆弱性辨识方法

为辨识道路网络脆弱性,引入道路网络鲁棒性指数评价方法,建立了考虑路段排队容量约束和断面通行能力约束的网络交通流模型,并利用Lagrange对偶算法将有约束问题转化为无约束问题求解了该模型。基于所建模型,通过依次断裂交通网络各路段以模拟路网受损,评价各断裂路段对应的道路网络鲁棒性指数,进而依据指数的排序结果辨识了道路网络的脆弱性。结果表明:基于网络鲁棒性指数的辨识方法可以克服基于饱和度的辨识方法的缺陷,能从全局角度准确辨识道路网络脆弱性;辨识模型中是否考虑路段排队容量和断面通行能力约束影响着道路网络脆弱性辨识结果的准确性;提出的基于鲁棒性指数的考虑路段排队容量约束和路段断面通行能力约束的辨识模型,能准确识别导致道路网络脆弱性显著的关键路段,这将为路段抢修资源分配和道路网络改造优先时序决策提供支持。