Delay optimization of unsignalized intersections based on petri nets and an improved exponential distribution optimization

For the scheduling problem of vehicles at unsignalized intersections, this paper proposes an improved exponential distribution optimization (IEDO) by a Petri net model and a deadlock avoidance policy with minimizing the delay of the system. Firstly, encoding is presented between vehicle numbers and EDO individuals; Secondly, deadlock avoidance strategy is conducted on individuals to guarantee that the vehicle passage sequence is a feasible sequence; Next, three enhancements of IEDO, standardized mutation, stochastic inertia weights and metropolis criterion are introduced to improve the algorithm’s optimal search capability. Finally, the effectiveness of IEDO is verified by experiments on the two-way four-lane intersection. After experimental comparisons, it is verified that IEDO has better optimization seeking capability in solving the vehicle scheduling problem at unsignalized intersections.