Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation
Since simulation-based optimisation typically requires large numbers of runs to identify sufficiently good solutions, the costs in terms of time and hardware can be enormous. To avoid unnecessary simulation runs, surrogate models can be applied, which estimate the simulation output under a given par...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/143197 |
| _version_ | 1811693806727200768 |
|---|---|
| author | Andelfinger, Philip Udayakumar, Sajeev Cai, Wentong Eckhoff, David Knoll, Alois |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Andelfinger, Philip Udayakumar, Sajeev Cai, Wentong Eckhoff, David Knoll, Alois |
| author_sort | Andelfinger, Philip |
| collection | NTU |
| description | Since simulation-based optimisation typically requires large numbers of runs to identify sufficiently good solutions, the costs in terms of time and hardware can be enormous. To avoid unnecessary simulation runs, surrogate models can be applied, which estimate the simulation output under a given parameter combination. Model preemption is a related technique that dynamically analyses the simulation state at runtime to identify runs unlikely to result in a high-quality solution and terminates such runs early. However, existing work on model preemption relies on model-specific termination rules. In this paper, we describe an architecture for simulation-based optimisation using model preemption based on estimations of the simulation output. In a case study, the approach is applied to the optimisation of traffic light timings in a traffic simulation. We show that within a given time and hardware budget, model preemption enables the identification of higher-quality solutions than those found through traditional simulation-based optimisation. |
| first_indexed | 2024-10-01T06:57:32Z |
| format | Conference Paper |
| id | ntu-10356/143197 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:57:32Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | ntu-10356/1431972020-08-12T04:39:52Z Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation Andelfinger, Philip Udayakumar, Sajeev Cai, Wentong Eckhoff, David Knoll, Alois School of Computer Science and Engineering 2018 Winter Simulation Conference (WSC) TUMCREATE Ltd. Engineering::Computer science and engineering Optimization Computational Modeling Since simulation-based optimisation typically requires large numbers of runs to identify sufficiently good solutions, the costs in terms of time and hardware can be enormous. To avoid unnecessary simulation runs, surrogate models can be applied, which estimate the simulation output under a given parameter combination. Model preemption is a related technique that dynamically analyses the simulation state at runtime to identify runs unlikely to result in a high-quality solution and terminates such runs early. However, existing work on model preemption relies on model-specific termination rules. In this paper, we describe an architecture for simulation-based optimisation using model preemption based on estimations of the simulation output. In a case study, the approach is applied to the optimisation of traffic light timings in a traffic simulation. We show that within a given time and hardware budget, model preemption enables the identification of higher-quality solutions than those found through traditional simulation-based optimisation. National Research Foundation (NRF) Accepted version This work was financially supported by the Singapore National Research Foundation under its Campus for Research Excellence And Technological Enterprise (CREATE) programme. 2020-08-12T04:39:52Z 2020-08-12T04:39:52Z 2019 Conference Paper Andelfinger, P., Udayakumar, S., Cai, W., Eckhoff, D., & Knoll, A. (2018). Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation. Proceedings of the 2018 Winter Simulation Conference (WSC), 652-663. doi:10.1109/WSC.2018.8632411 978-1-5386-6573-2 https://hdl.handle.net/10356/143197 10.1109/WSC.2018.8632411 2-s2.0-85062640550 652 663 en © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/WSC.2018.8632411. application/pdf |
| spellingShingle | Engineering::Computer science and engineering Optimization Computational Modeling Andelfinger, Philip Udayakumar, Sajeev Cai, Wentong Eckhoff, David Knoll, Alois Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title | Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title_full | Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title_fullStr | Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title_full_unstemmed | Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title_short | Model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| title_sort | model preemption based on dynamic analysis of simulation data to accelerate traffic light timing optimisation |
| topic | Engineering::Computer science and engineering Optimization Computational Modeling |
| url | https://hdl.handle.net/10356/143197 |
| work_keys_str_mv | AT andelfingerphilip modelpreemptionbasedondynamicanalysisofsimulationdatatoacceleratetrafficlighttimingoptimisation AT udayakumarsajeev modelpreemptionbasedondynamicanalysisofsimulationdatatoacceleratetrafficlighttimingoptimisation AT caiwentong modelpreemptionbasedondynamicanalysisofsimulationdatatoacceleratetrafficlighttimingoptimisation AT eckhoffdavid modelpreemptionbasedondynamicanalysisofsimulationdatatoacceleratetrafficlighttimingoptimisation AT knollalois modelpreemptionbasedondynamicanalysisofsimulationdatatoacceleratetrafficlighttimingoptimisation |