Trafik sinyal optimizasyonu için çok amaçlı doğrusal programlama yaklaşımlarının geliştirilmesi

In many countries and metropolitan cities, traffic congestion, mainly caused by population growth and the increase in urbanization, has reached a significant level and has become a major problem for residents and decision-makers. Today, although infrastructural strategies such as the construction of underpasses and road widening to alleviate traffic congestion are applied in urban networks, the strategies can also be quite costly and environmentally damaging. Therefore, decision makers allocate large budgets to solve transportation problems and alleviate traffic congestion in urban road networks. In addition, in some cases, infrastructural strategies (road widening, additional lanes, underpasses etc.) can not always be possible due to environmental factors. For this reason, many studies and research have been carried out over the last 40-50 years with the aim of designing and developing effective traffic signal control strategies, which are the most cost-effective and economical solution to the problem of traffic congestion in urban road networks. In this study, various linear programming approaches with different objective functions (fair allocation of residual queue, minimization of total cycle length, minimization of total residual queue, etc.) and fixed-time traffic signal control strategies have been developed, aiming to reduce traffic congestion according to various traffic demands traffic at isolated signalized intersections. In under-saturated conditions, MCLM and CCM linear programming approaches were developed and the green times optimized by the approaches were compared in terms of delay, which is the most important performance criteria using HCM 2000 delay model. In over-saturated conditions, on the other hand, a 2-stage approach has been developed. In the first stage, two different linear programming approaches, MTQLM and MMQLM, were created and the green times of each approach that gave low average vehicle delay according to the traffic scenario were obtained in the first stage with the HCM 2000 delay calculation. In the second stage, the green times obtained in the first scene were optimized to find new candidate green times using HCM 2000 delay model, with the help of neighbor search algorithm (NSM) in a predetermined range. The developed models were evaluated at 3 different types of intersection (T-Type, 4-Legged) with different intersection geometries. The models are compared according to the HCM 2000 average vehicle delay and the reasons for the effects of the models on the delays are also discussed. In addition, in order to measure how generated approaches behave in real-life, the 3rd intersection was also analyzed in the PTV VISSIM microsimulation environment and the delays obtained in the PTV VISSIM environment were compared with HCM 2000.