On-demand Ride-pooling with Walking Legs: Decomposition Approach for Dynamic Matching and Virtual Stops Selection. Sarma, N. J., Gurumurthy, K. M., Hyland, M., Bahk, Y., de Souza, F., Verbas, O., & Wang, Z. In Washington, D.C., August, 2023. abstract bibtex Door-to-door (D2D) ride-hailing services currently dominate the mobility-on-demand (MOD) market, but several alternative MOD service types offer operational and societal benefits. Three such MOD services include D2D ride-pooling, corner-to-corner (C2C) ride-hailing, and C2C ride-pooling. C2C service requires travelers to walk a short distance to/from a pickup/drop-off location. The goals of this study are two-fold. First, we compare these four MOD services in terms of operator costs (e.g., vehicle kilometers per request served) and user costs (e.g., assignment time, wait time, walk time, and invehicle time). Second, we develop an effective and scalable decision policy and solution algorithm for operating a C2C ride-pooling service. At each decision epoch, the operator must dynamically assign vehicles to requests, route and schedule vehicles, and assign travelers to pickup and drop-off (PUDO) locations. To address this problem, we propose decomposing the problem into a matching, routing, and scheduling subproblem, and a PUDO locations selection subproblem. We use geographic, network, and vehicle information, as well as optimization techniques to solve the two subproblems. The computational experiments confirm a clear trade-off across the four services in terms of operator costs and user costs. With D2D ridehailing as the baseline, (i) ride-pooling significantly reduces operator costs, while slightly increasing user costs; (ii) C2C slightly reduces operator costs while increasing user costs; (iii) combining ride-pooling and C2C appears to provide additive benefits in terms of operator costs.
@inproceedings{sarma_-demand_2023,
address = {Washington, D.C.},
title = {On-demand {Ride}-pooling with {Walking} {Legs}: {Decomposition} {Approach} for {Dynamic} {Matching} and {Virtual} {Stops} {Selection}},
abstract = {Door-to-door (D2D) ride-hailing services currently dominate the mobility-on-demand (MOD) market, but several alternative MOD service types offer operational and societal benefits. Three such MOD services include D2D ride-pooling, corner-to-corner (C2C) ride-hailing, and C2C ride-pooling. C2C service requires travelers to walk a short distance to/from a pickup/drop-off location. The goals of this study are two-fold. First, we compare these four MOD services in terms of operator costs (e.g., vehicle kilometers per request served) and user costs (e.g., assignment time, wait time, walk time, and invehicle time). Second, we develop an effective and scalable decision policy and solution algorithm for operating a C2C ride-pooling service. At each decision epoch, the operator must dynamically assign vehicles to requests, route and schedule vehicles, and assign travelers to pickup and drop-off (PUDO) locations. To address this problem, we propose decomposing the problem into a matching, routing, and scheduling subproblem, and a PUDO locations selection subproblem. We use geographic, network, and vehicle information, as well as optimization techniques to solve the two subproblems. The computational experiments confirm a clear trade-off across the four services in terms of operator costs and user costs. With D2D ridehailing as the baseline, (i) ride-pooling significantly reduces operator costs, while slightly increasing user costs; (ii) C2C slightly reduces operator costs while increasing user costs; (iii) combining ride-pooling and C2C appears to provide additive benefits in terms of operator costs.},
author = {Sarma, Navjyoth J.S and Gurumurthy, Krishna M. and Hyland, Michael and Bahk, Younghun and de Souza, Felipe and Verbas, Omer and Wang, Zifan},
month = aug,
year = {2023},
}
Downloads: 0
{"_id":"fezkEyGwnCrcKDGgH","bibbaseid":"sarma-gurumurthy-hyland-bahk-desouza-verbas-wang-ondemandridepoolingwithwalkinglegsdecompositionapproachfordynamicmatchingandvirtualstopsselection-2023","author_short":["Sarma, N. J.","Gurumurthy, K. M.","Hyland, M.","Bahk, Y.","de Souza, F.","Verbas, O.","Wang, Z."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Washington, D.C.","title":"On-demand Ride-pooling with Walking Legs: Decomposition Approach for Dynamic Matching and Virtual Stops Selection","abstract":"Door-to-door (D2D) ride-hailing services currently dominate the mobility-on-demand (MOD) market, but several alternative MOD service types offer operational and societal benefits. Three such MOD services include D2D ride-pooling, corner-to-corner (C2C) ride-hailing, and C2C ride-pooling. C2C service requires travelers to walk a short distance to/from a pickup/drop-off location. The goals of this study are two-fold. First, we compare these four MOD services in terms of operator costs (e.g., vehicle kilometers per request served) and user costs (e.g., assignment time, wait time, walk time, and invehicle time). Second, we develop an effective and scalable decision policy and solution algorithm for operating a C2C ride-pooling service. At each decision epoch, the operator must dynamically assign vehicles to requests, route and schedule vehicles, and assign travelers to pickup and drop-off (PUDO) locations. To address this problem, we propose decomposing the problem into a matching, routing, and scheduling subproblem, and a PUDO locations selection subproblem. We use geographic, network, and vehicle information, as well as optimization techniques to solve the two subproblems. The computational experiments confirm a clear trade-off across the four services in terms of operator costs and user costs. With D2D ridehailing as the baseline, (i) ride-pooling significantly reduces operator costs, while slightly increasing user costs; (ii) C2C slightly reduces operator costs while increasing user costs; (iii) combining ride-pooling and C2C appears to provide additive benefits in terms of operator costs.","author":[{"propositions":[],"lastnames":["Sarma"],"firstnames":["Navjyoth","J.S"],"suffixes":[]},{"propositions":[],"lastnames":["Gurumurthy"],"firstnames":["Krishna","M."],"suffixes":[]},{"propositions":[],"lastnames":["Hyland"],"firstnames":["Michael"],"suffixes":[]},{"propositions":[],"lastnames":["Bahk"],"firstnames":["Younghun"],"suffixes":[]},{"propositions":["de"],"lastnames":["Souza"],"firstnames":["Felipe"],"suffixes":[]},{"propositions":[],"lastnames":["Verbas"],"firstnames":["Omer"],"suffixes":[]},{"propositions":[],"lastnames":["Wang"],"firstnames":["Zifan"],"suffixes":[]}],"month":"August","year":"2023","bibtex":"@inproceedings{sarma_-demand_2023,\n\taddress = {Washington, D.C.},\n\ttitle = {On-demand {Ride}-pooling with {Walking} {Legs}: {Decomposition} {Approach} for {Dynamic} {Matching} and {Virtual} {Stops} {Selection}},\n\tabstract = {Door-to-door (D2D) ride-hailing services currently dominate the mobility-on-demand (MOD) market, but several alternative MOD service types offer operational and societal benefits. Three such MOD services include D2D ride-pooling, corner-to-corner (C2C) ride-hailing, and C2C ride-pooling. C2C service requires travelers to walk a short distance to/from a pickup/drop-off location. The goals of this study are two-fold. First, we compare these four MOD services in terms of operator costs (e.g., vehicle kilometers per request served) and user costs (e.g., assignment time, wait time, walk time, and invehicle time). Second, we develop an effective and scalable decision policy and solution algorithm for operating a C2C ride-pooling service. At each decision epoch, the operator must dynamically assign vehicles to requests, route and schedule vehicles, and assign travelers to pickup and drop-off (PUDO) locations. To address this problem, we propose decomposing the problem into a matching, routing, and scheduling subproblem, and a PUDO locations selection subproblem. We use geographic, network, and vehicle information, as well as optimization techniques to solve the two subproblems. The computational experiments confirm a clear trade-off across the four services in terms of operator costs and user costs. With D2D ridehailing as the baseline, (i) ride-pooling significantly reduces operator costs, while slightly increasing user costs; (ii) C2C slightly reduces operator costs while increasing user costs; (iii) combining ride-pooling and C2C appears to provide additive benefits in terms of operator costs.},\n\tauthor = {Sarma, Navjyoth J.S and Gurumurthy, Krishna M. and Hyland, Michael and Bahk, Younghun and de Souza, Felipe and Verbas, Omer and Wang, Zifan},\n\tmonth = aug,\n\tyear = {2023},\n}\n\n","author_short":["Sarma, N. J.","Gurumurthy, K. M.","Hyland, M.","Bahk, Y.","de Souza, F.","Verbas, O.","Wang, Z."],"key":"sarma_-demand_2023","id":"sarma_-demand_2023","bibbaseid":"sarma-gurumurthy-hyland-bahk-desouza-verbas-wang-ondemandridepoolingwithwalkinglegsdecompositionapproachfordynamicmatchingandvirtualstopsselection-2023","role":"author","urls":{},"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/2254223/items?key=Bbol5AZSx3A5fDcmq91mfOKd&format=bibtex&limit=100","dataSources":["8NTJjnAGyZ5qTfMug"],"keywords":[],"search_terms":["demand","ride","pooling","walking","legs","decomposition","approach","dynamic","matching","virtual","stops","selection","sarma","gurumurthy","hyland","bahk","de souza","verbas","wang"],"title":"On-demand Ride-pooling with Walking Legs: Decomposition Approach for Dynamic Matching and Virtual Stops Selection","year":2023}