TY - CHAP
T1 - Heuristic algorithm for solving a multimodal location-based concierge service problem
AU - Kang, Seungmo
AU - Oh, Seung
AU - Kim, Tschangho John
PY - 2006
Y1 - 2006
N2 - As a location-based services problem, the concierge service problem is to find a minimum total cost for purchasing predetermined types and quantities of items with the shortest paths connecting the locations where the items are available. These locations are defined as points of interest (POIs). The total cost includes purchasing and stopping costs at POIs as well as the travel cost from origin to destination. To respond to a request for such a service, a search algorithm should be reasonably fast in computational time and high in accuracy. A heuristic search algorithm was developed by a Euclidean distance approach using a geographic information system. The algorithm was implemented with 1,248 POIs in the Seoul, South Korea, metropolitan area as a case study site. The road network is composed of 52,915 nodes and 77,339 links, and the existing subway network, and all existing bus routes were used for the implementation. Four scenarios were examined: a request for a service in peak and off-peak periods with and without turning restrictions. The results indicate that the computational time in each case is less than 4 s with the transit networks only and less than 30 s if the road network is used. The paper also evaluates the solutions generated from the algorithm.
AB - As a location-based services problem, the concierge service problem is to find a minimum total cost for purchasing predetermined types and quantities of items with the shortest paths connecting the locations where the items are available. These locations are defined as points of interest (POIs). The total cost includes purchasing and stopping costs at POIs as well as the travel cost from origin to destination. To respond to a request for such a service, a search algorithm should be reasonably fast in computational time and high in accuracy. A heuristic search algorithm was developed by a Euclidean distance approach using a geographic information system. The algorithm was implemented with 1,248 POIs in the Seoul, South Korea, metropolitan area as a case study site. The road network is composed of 52,915 nodes and 77,339 links, and the existing subway network, and all existing bus routes were used for the implementation. Four scenarios were examined: a request for a service in peak and off-peak periods with and without turning restrictions. The results indicate that the computational time in each case is less than 4 s with the transit networks only and less than 30 s if the road network is used. The paper also evaluates the solutions generated from the algorithm.
UR - http://www.scopus.com/inward/record.url?scp=33947610809&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33947610809&partnerID=8YFLogxK
U2 - 10.3141/1972-17
DO - 10.3141/1972-17
M3 - Chapter
AN - SCOPUS:33947610809
SN - 0309099811
SN - 9780309099813
T3 - Transportation Research Record
SP - 123
EP - 132
BT - Travel Survey Methods, Information Technology, and Geospatial Data
PB - National Research Council
ER -