In the proposed research, we will apply statistical physics to understand the nature and the limitation of transportation optimization, and use these insights to derive practical optimization algorithms. Similar success by statistical physics has been demonstrated in other optimization problems, which has led to ground-breaking advances. Our objective is threefold. Firstly, we aim to reveal the dynamics and the interplay of routing strategies leading to user equilibriums. We then formulate a simple model to understand analytically the emergence of these sub-optimal states. Secondly, we aim to devise algorithms which coordinate the spatial-temporal routes of individuals, driving the system towards the global optimum. We will also reveal the density of sub-optimal states in the state space, which lead to insights into the intrinsic sub-optimality of transportation networks and thus the limitation of optimization algorithms. Finally, we aim to devise algorithms to optimally divert traffic in cases of disturbances, e.g. road blockage due car crashes, which are less explored but as important as recurrent traffic optimization.
|Effective start/end date||01/01/16 → 30/06/19|