Optimization of Traffic Signal Settings in Networks by Mixed-Integer Linear Programming
MASSACHUSETTS INST OF TECH CAMBRIDGE OPERATIONS RESEARCH CENTER
Pagination or Media Count:
A mixed-integer linear programming formulation is developed for minimizing delay to traffic in a signal controlled road network. Offsets, splits of green time and a common cycle time for the network are considered as decision variables simultaneously. The traffic flow pattern is modeled as a periodic platoon, and a link performance function is derived in the form of a piecewise linear convex surface representing the delay incurred by these platoons. Stochastic effects are accounted for by a saturation deterrence function representing the expected overflow queue on each link and are included as an additive component in the objective function. Computational results, using the MPSX system, are given for an arterial with 11 signals in Waltham, Mass., and a portion of the UTCS network in Washington, D.C. containing 20 nodes, 63 links and 21 loops.
- Operations Research
- Civil Engineering