INTERAKSI ANTARA PEMBEBANAN LALU LINTAS DAN PENGENDALIAN SINYAL LALU LINTAS

• Main Author: Hartawan Tjahja, Vici
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/2949
• Institution: Institut Teknologi Bandung

<br /> Traffic assignment and signal control are generally considered separately because standard signal setting policies commonly used are based on the fundamental assumption that route choices are unaffected by the signal settings chosen. This is reasonable when route-choice is obvious and severely limited, or congestion is slight, but it is not reasonable in general. Usually, no account is taken either of the effect of these signal settings on the future routing decisions of drivers, or of the effect any such re-routing may have on the efficiency of the network as a whole. It is therefore natural to seek signal setting policies which are optimal when route-choice decisions by drivers are taken into account. The need to effectively integrate these two perspectives has motivated the development over the last two decades of a class of models known as combined traffic assignment and control (CTAC) models. However, implementation of these models and comparative tests of the performance of alternative signal control policies that have been published were very limited. This research was designed to compare and investigate the performances of two standard responsive control policies, i.e. equisaturation policy similar to Webster\'s and the familiar local delay-minimising policy, with the performance of the specially-designed policy Po, allowing for the likely future re-routing decisions of drivers. To allow for future re-routing in the evaluation of long-run network performances, these policies were compared within a steady-state assignment-control model. Four artificially-constructed networks and one realistic network within City of Bandung were examined in this study. It was difficult to make a recommendation for a policy to be used, as the performance of the three policies varied with the network. From tests conducted using various networks, it was found that the performance of a control policy is subject to the network configuration and the trip demand level. However, test results from the real network showed that a better traffic condition could be obtained by using a CTAC model