Incremental multistep Q-learning for adaptive traffic signal control based on delay minimization strategy

Incremental multistep Q learning (Q( λ )) combines Q learning and TD(λ ). Theoretically, Q(λ ) has better performance than Q learning. The goal of the paper is to test the performance of Q(λ ) for adaptive traffic signal control. For Q(λ ), the state is total delay of the intersection, and the action is phase green time change. The relationship between phase green time change and action space is discussed. The performance between Q(λ) learning and fixed cycle signal setting for isolated intersection is compared. The computation results show that Q(λ ) learning for traffic signal control can achieve lesser delay for variable traffic condition.