00000ctm a22000004i 4500 UP-99796217613096951 Buklod 20230215105418.0 m |o d | ta 190905s xx d r |||| u| (iLib)UPD-00413166081 DENGII rda eng DMLUC LG 993.5 2019 C65 D45 Dela Cruz, Francis Zac author. A cell-based multi-objective approach to traffic signal optimization Francis Zac Dela Cruz and Luis Miguel Flores ; John Justine S. Villar, adviser ; Anjilo M. Carigma, co-adviser. Quezon City College of Engineering, University of the Philippines Diliman 2019. x, 34 leaves color illustrations 28 cm text rdacontent unmediated rdamedia volume rdacarrier Thesis (Bachelor of Science in Computer Science)--University of the Philippines Diliman May 2019. F - no patentable invention or creation, not for personal publication and no confidential information. Yes - available to the general public. The paper proposes a multi-objective approach to optimizing the signal timing of traffic lights in a single four-legged intersection in real time. The model considers delay minimization and throughput maximization in its objective function, and allocates weights to allow flexibility on which term is given more importance based on the state of the intersection. The vehicle holding problem is also addressed by the model. Seven weight configurations were sampled in order to observe the performance of the model with different weight allocations. Four intersection states, each with varying demands, were considered in the testing of the model. A mixed-integer linear program, or MILP, was used to solve for optimal greentimes every three cycles, which were fed into a microsimulator for analysis. The proposed approach was compared to an existing model that focused solely on throughput, and their performance was measured with the resulting runtime, average delay, travel time, and total throughput. The proposed model was able to recompute greentimes much faster than the parent model, allowing for more real-time applications. It was also able to outperform the parent model in terms of average delay, and was able to equally distribute travel time between left and through movements among approaches. The parent model performed better in terms of throughput, but only within 9%. For future research, the ring barrier configuration should be revisited for both models. Since the ring barrier configurations used in the parent and proposed models differ, the disparity in their performance may be attributed to that difference. Traffic signs and signals Mathematical models. Flores, Luis Miguel author. Villar, John Justine S. adviser. Carigma, Anjilo M. adviser. Thesis FI UP UPD DENG-II LG 993.5 2019 C65 D45 Thesis