Stochastic Gradient-Based Optimal Signal Control With Energy Consumption Bounds

摘要: This paper develops a stochastic gradient-based optimization model for traffic signal control with bounds on network-level vehicular energy consumption. The problem is formulated as mixed-integer linear mathematical program, which incorporates inequality constraints to limit the total consumption in network. developed gradient approximation algorithm provides near-optimal solution non-convex problem. To account constraints, penalty function method leveraging pseudo estimation technique developed. Empirical results from signalized arterial street show that it possible achieve optimized settings at desired bound without compromising delay. Further, we report sensitivity of mobility metrics—system Our novel gradient-approximation-based offers functional and feasible way accommodate within maximal minimal