Urban traffic resilience control – An ecological resilience perspective
arxiv(2024)
摘要
Urban traffic resilience has gained increased attention, with most studies
adopting an engineering perspective that assumes a single optimal equilibrium
and prioritizes local recovery. On the other hand, systems may possess multiple
metastable states, and ecological resilience is the ability to switch between
these states according to perturbations. Control strategies from these two
resilience perspectives yield distinct outcomes. In fact, ecological resilience
oriented control has rarely been viewed in urban traffic, despite the fact that
traffic system is a complex system in highly uncertain environment with
possible multiple metastable states. This absence highlights the necessity for
urban traffic ecological resilience definition. To bridge this gap, we defines
urban traffic ecological resilience as the ability to absorb uncertain
perturbations by shifting to alternative states. The goal is to generate a
system with greater adaptability, without necessarily returning to the original
equilibrium. Our control framework comprises three aspects: portraying the
recoverable scopes; designing alternative steady states; and controlling system
to shift to alternative steady states for adapting large disturbances. Among
them, the recoverable scopes are portrayed by attraction region; the
alternative steady states are set close to the optimal state and outside the
attraction region of the original equilibrium; the controller needs to ensure
the local stability of the alternative steady states, without changing the
trajectories inside the attraction region of the original equilibrium.
Comparisons with classical engineering resilience oriented urban traffic
resilience control schemes show that, proposed ecological resilience oriented
control schemes can generate greater resilience. These results will contribute
to the fundamental theory of future resilient intelligent transportation
system.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要