Krill Herd Algorithm for Signal Optimization of Cooperative Control With Traffic Supply and Demand

Abstract

As a novel signal control method, signal Cooperative Control with traffic Supply and Demand (CCSD) is superior to the traditional control methods and could satisfy the control requirements under all traffic conditions. However, the optimization solution of CCSD cannot meet the real-time control requirement for its exhaustive search. To overcome the problem, in this paper, the optimization problem of CCSD is reconstructed by the time-varying traffic supply and demand, and the problem is reduced as a problem of space search. Then, the krill herd (KH) algorithm is introduced and employed to realize the fast solution of CCSD (named KH-CCSD). During the process of optimization, the search space representation and fitness function for the KH algorithm are constructed to satisfy the solution of CCSD. The optimal signal timing plan is obtained by an iterated search of krill swarm in a multi-dimensional time-varying space cooperatively constrained by traffic supplies and demands. The convergence and effectiveness of KH-CCSD are validated by comparing experiments, in which the convergence of KH-CCSD is tested by different initializations and KH-CCSD is compared with the Webster method and capacity-aware back-pressure (CABP) control under unsaturated, saturated, and oversaturated conditions. The experiments results show that KH-CCSD performs a fast convergence and KH-CCSD is superior to the Webster method and CABP. As a result, KH-CCSD could satisfy the application of CCSD under all traffic conditions.