A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation

博士 === 國立交通大學 === 資訊科學與工程研究所 === 106 === Routing has many applications in industry and life. In industry, routing is a main step in the design of integrated circuits (ICs). Multi-layer obstacle-avoiding rectilinear Steiner tree (ML-OARST) construction is crucial for a router or a placer to approxima...

Full description

Bibliographic Details
Main Authors: Lin, Kuen-Wey, 林昆蔚
Other Authors: Li, Yih-Lang
Format: Others
Language:en_US
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/ee82jd
id ndltd-TW-106NCTU5394162
record_format oai_dc
collection NDLTD
language en_US
format Others
sources NDLTD
description 博士 === 國立交通大學 === 資訊科學與工程研究所 === 106 === Routing has many applications in industry and life. In industry, routing is a main step in the design of integrated circuits (ICs). Multi-layer obstacle-avoiding rectilinear Steiner tree (ML-OARST) construction is crucial for a router or a placer to approximate or simply complete the routing of a net. In life, a routing algorithm embedded in an on-board global positioning system (GPS) device is used to navigate a vehicle to avoid traffic congestion and reduce driving time. Because it is difficult to find empty space in a developed city to accommodate more transportation infrastructures, the development of an effective routing algorithm for navigation is a low cost option for mitigating traffic jam. In this dissertation, the minimum Steiner tree algorithm is addressed, and an idea of retention of routing information is proposed for the applications of routing in ML-OARST and navigation. For ML-OARST, the idea is utilized to select good Steiner points to reduce the routing cost of a tree; for navigation, the idea is implemented with timetable to evaluate near-future traffic information and mitigate traffic jam accordingly. For the applications of routing in the ML-OARST construction, a maze routing-based methodology with bounded exploration and path-assessed retracing is presented. The idea of bounded maze routing not only reduces the propagation of maze routing but also reserves good candidates for Steiner points. A method that conveniently assesses the paths of connecting pins in the retracing phase of maze routing was developed to make full use of the features of maze routing. The retention of path assessment is used to select good Steiner points. The refinement of tree topology is implemented with ripping up and re-building the set of Steiner points instead of traditional edge/vertex removal or edge shifting/substitution. Layer directive and weighted layer are also considered. A commercial tool is used to show the effectiveness of the proposed methodology. Moreover, for the construction of a performance-driven (PD) ML-OARST, the dissertation presents an algorithm of PD minimum-cost spanning tree construction to connect the set of pins and the set Steiner points, and a refinement for the PD metrics of the tree is also provided. Based on real-time traffic information, a GPS device or navigation service provider can find a locally optimal route for currently issued query, where a locally optimal route means the optimality is identified at a certain point in time and might change as time goes forward. To overcome this limitation, a dynamic update scheme can give a quick response to traffic jams as soon as possible by dynamically offering drivers another better route especially as vehicles slow down or even stop due to serious traffic jams. However the detour to avoid traffic jams is usually suggested as any one road included in the pre-planned route is going to be congested, and at that moment drivers might have passed some redundant paths that lead to congested roads or escape away from congested roads. This dissertation presents a routing system using timetables to implement the idea of retention of routing information. A timetable is utilized to store the present traffic information and evaluate the near-future traffic conditions of a road. Regarding a future world where automated driving technologies have become mature and people also get used to taking automated driving vehicles as transportation, the scheduled route of each vehicle is known to navigation system; thus, the proposed routing system owns a unique advantage of accurate traffic-jam evaluation over traditional navigation system. The routes of all vehicles can be evaluated by the proposed routing system, and route distribution and preventing near-future traffic jams can also be realized. To get close to the real world, some kinds of uncertainties are allowed, such as abrupt traffic jams caused by vehicle accident and prediction failure caused by changing the destination or a temporary stop. Experimental results with actual city maps show the evaluated near-future traffic information is helpful to mitigate traffic jam and reduce driving time.
author2 Li, Yih-Lang
author_facet Li, Yih-Lang
Lin, Kuen-Wey
林昆蔚
author Lin, Kuen-Wey
林昆蔚
spellingShingle Lin, Kuen-Wey
林昆蔚
A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
author_sort Lin, Kuen-Wey
title A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
title_short A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
title_full A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
title_fullStr A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
title_full_unstemmed A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation
title_sort study on minimum steiner tree algorithms and the application of routing algorithms to near-future traffic evaluation based automatic navigation
publishDate 2018
url http://ndltd.ncl.edu.tw/handle/ee82jd
work_keys_str_mv AT linkuenwey astudyonminimumsteinertreealgorithmsandtheapplicationofroutingalgorithmstonearfuturetrafficevaluationbasedautomaticnavigation
AT línkūnwèi astudyonminimumsteinertreealgorithmsandtheapplicationofroutingalgorithmstonearfuturetrafficevaluationbasedautomaticnavigation
AT linkuenwey zuìxiǎoshǐtǎnnàshùyǎnsuànfǎyǔràoxiànyǎnsuànfǎyīngyòngyúzàijùyǒuyùcèwèiláijiāotōngpíngjǐngyǔdòngtàidiàozhěngjiàshǐlùjìngnénglìdezìdòngdǎohángxìtǒngzhīyánjiū
AT línkūnwèi zuìxiǎoshǐtǎnnàshùyǎnsuànfǎyǔràoxiànyǎnsuànfǎyīngyòngyúzàijùyǒuyùcèwèiláijiāotōngpíngjǐngyǔdòngtàidiàozhěngjiàshǐlùjìngnénglìdezìdòngdǎohángxìtǒngzhīyánjiū
AT linkuenwey studyonminimumsteinertreealgorithmsandtheapplicationofroutingalgorithmstonearfuturetrafficevaluationbasedautomaticnavigation
AT línkūnwèi studyonminimumsteinertreealgorithmsandtheapplicationofroutingalgorithmstonearfuturetrafficevaluationbasedautomaticnavigation
_version_ 1719175843542990848
spelling ndltd-TW-106NCTU53941622019-05-16T01:24:32Z http://ndltd.ncl.edu.tw/handle/ee82jd A Study on Minimum Steiner Tree Algorithms and the Application of Routing Algorithms to Near-Future Traffic Evaluation Based Automatic Navigation 最小史坦納樹演算法與繞線演算法應用於在具有預測未來交通瓶頸與動態調整駕駛路徑能力的自動導航系統之研究 Lin, Kuen-Wey 林昆蔚 博士 國立交通大學 資訊科學與工程研究所 106 Routing has many applications in industry and life. In industry, routing is a main step in the design of integrated circuits (ICs). Multi-layer obstacle-avoiding rectilinear Steiner tree (ML-OARST) construction is crucial for a router or a placer to approximate or simply complete the routing of a net. In life, a routing algorithm embedded in an on-board global positioning system (GPS) device is used to navigate a vehicle to avoid traffic congestion and reduce driving time. Because it is difficult to find empty space in a developed city to accommodate more transportation infrastructures, the development of an effective routing algorithm for navigation is a low cost option for mitigating traffic jam. In this dissertation, the minimum Steiner tree algorithm is addressed, and an idea of retention of routing information is proposed for the applications of routing in ML-OARST and navigation. For ML-OARST, the idea is utilized to select good Steiner points to reduce the routing cost of a tree; for navigation, the idea is implemented with timetable to evaluate near-future traffic information and mitigate traffic jam accordingly. For the applications of routing in the ML-OARST construction, a maze routing-based methodology with bounded exploration and path-assessed retracing is presented. The idea of bounded maze routing not only reduces the propagation of maze routing but also reserves good candidates for Steiner points. A method that conveniently assesses the paths of connecting pins in the retracing phase of maze routing was developed to make full use of the features of maze routing. The retention of path assessment is used to select good Steiner points. The refinement of tree topology is implemented with ripping up and re-building the set of Steiner points instead of traditional edge/vertex removal or edge shifting/substitution. Layer directive and weighted layer are also considered. A commercial tool is used to show the effectiveness of the proposed methodology. Moreover, for the construction of a performance-driven (PD) ML-OARST, the dissertation presents an algorithm of PD minimum-cost spanning tree construction to connect the set of pins and the set Steiner points, and a refinement for the PD metrics of the tree is also provided. Based on real-time traffic information, a GPS device or navigation service provider can find a locally optimal route for currently issued query, where a locally optimal route means the optimality is identified at a certain point in time and might change as time goes forward. To overcome this limitation, a dynamic update scheme can give a quick response to traffic jams as soon as possible by dynamically offering drivers another better route especially as vehicles slow down or even stop due to serious traffic jams. However the detour to avoid traffic jams is usually suggested as any one road included in the pre-planned route is going to be congested, and at that moment drivers might have passed some redundant paths that lead to congested roads or escape away from congested roads. This dissertation presents a routing system using timetables to implement the idea of retention of routing information. A timetable is utilized to store the present traffic information and evaluate the near-future traffic conditions of a road. Regarding a future world where automated driving technologies have become mature and people also get used to taking automated driving vehicles as transportation, the scheduled route of each vehicle is known to navigation system; thus, the proposed routing system owns a unique advantage of accurate traffic-jam evaluation over traditional navigation system. The routes of all vehicles can be evaluated by the proposed routing system, and route distribution and preventing near-future traffic jams can also be realized. To get close to the real world, some kinds of uncertainties are allowed, such as abrupt traffic jams caused by vehicle accident and prediction failure caused by changing the destination or a temporary stop. Experimental results with actual city maps show the evaluated near-future traffic information is helpful to mitigate traffic jam and reduce driving time. Li, Yih-Lang Lin, Rung-Bin 李毅郎 林榮彬 2018 學位論文 ; thesis 80 en_US