Traffic Signal Optimization: Greedy Randomized Tabu Search Algorithm

摘要

Under Advanced Traffic Management Systems (ATMS), pre-timed signal control still plays an important role in traffic control and management. A wide variety of techniques have been proposed to generate optimal or near-optimal solutions for signal optimization problems. This research applies a randomized meta-heuristic algorithm, greedy randomized tabu search (GRTS), for network-level signal optimization problems. With the development of the algorithm, it can be a systematic approach for decision-making of traffic signal strategies to improve network performance. How to accurately evaluate the proposed algorithm is another issue tackled in this paper. Meta heuristic algorithms start searching procedure with generation a set of solution/s in the first iteration. Then simulation model of the considered problems is called and the fitness function is calculated for each individual to evaluate the goodness of such solution. In a same processing time, different algorithms may not end up with the same number of iterations and adversely, optimization algorithms even with the same number of iterations may not require the same computational time. In order to compare the performance of GRTS with other approaches, two signal optimization softwares (including Synchro and TRANSYT-7F) are also used to generate the optimal signal strategies. Numerical experiments on the test network are used in the comparison of GRTS algorithms and two software packages. A novel comparison criterion which is independent of computer capability and computational time is proposed and appropriate numerical comparisons are performed.