Discrete choice model acts as one of the most important tools for studies involving mode split in the context of transport demand forecast. As different types of discrete choice models display their merits and restric...Discrete choice model acts as one of the most important tools for studies involving mode split in the context of transport demand forecast. As different types of discrete choice models display their merits and restrictions diversely, how to properly select the specific type among discrete choice models for realistic application still remains to be a tough problem. In this article, five typical discrete choice models for transport mode split are, respectively, discussed, which includes multinomial logit model, nested logit model （NL）, heteroscedastic extreme value model, multinominal probit model and mixed multinomial logit model （MMNL）. The theoretical basis and application attributes of these five models are especially analysed with great attention, and they are also applied to a realistic intercity case of mode split forecast, which results indi- cating that NL model does well in accommodating similarity and heterogeneity across alternatives, while MMNL model serves as the most effective method for mode choice prediction since it shows the highest reliability with the least significant prediction errors and even outperforms the other four models in solving the heterogeneity and similarity problems. This study indicates that conclusions derived from a single discrete choice model are not reliable, and it is better to choose the proper model based on its characteristics.展开更多
In this work,we study the conic quadratic mixed-integer formulation for assortment optimization problem under the mixture of multinomial logit(MMNL)model.The MMNL model generalizes the widely studied multinomial logit...In this work,we study the conic quadratic mixed-integer formulation for assortment optimization problem under the mixture of multinomial logit(MMNL)model.The MMNL model generalizes the widely studied multinomial logit choice model and can approximate any random utility model with an arbitrary additive error.An important operational decision problem in revenue management is assortment optimization problem,which aims to find a subset of products to make available to customers that maximizes the expected revenue of the retailer.It is known that assortment optimization problem under the MMNL model is NP-hard and inapproximable within any constant performance guarantee.Commonly used methods for solving such problem are heuristical approaches or customized combinatorial optimization approaches.In the meanwhile,studies related to global optimization approaches are relatively scarce.We propose an enhanced conic quadratic mixed-integer formulation for solving assortment optimization problem under the MMNL model with a higher computational efficiency.Furthermore,we conduct extensive numerical experiments to demonstrate that the proposed reformulation significantly outperforms the existing conic reformulations for assortment optimization under the MMNL model.展开更多
基金supported by the Science&Technology pillar project(No.0556)of Guangzhou
文摘Discrete choice model acts as one of the most important tools for studies involving mode split in the context of transport demand forecast. As different types of discrete choice models display their merits and restrictions diversely, how to properly select the specific type among discrete choice models for realistic application still remains to be a tough problem. In this article, five typical discrete choice models for transport mode split are, respectively, discussed, which includes multinomial logit model, nested logit model （NL）, heteroscedastic extreme value model, multinominal probit model and mixed multinomial logit model （MMNL）. The theoretical basis and application attributes of these five models are especially analysed with great attention, and they are also applied to a realistic intercity case of mode split forecast, which results indi- cating that NL model does well in accommodating similarity and heterogeneity across alternatives, while MMNL model serves as the most effective method for mode choice prediction since it shows the highest reliability with the least significant prediction errors and even outperforms the other four models in solving the heterogeneity and similarity problems. This study indicates that conclusions derived from a single discrete choice model are not reliable, and it is better to choose the proper model based on its characteristics.
基金supported by the Fundamental Research Funds for the Central Universities of Xiamen University(No.2072021127)Ka-Meng Nip’s research work is partially supported by the Natural Science Foundation of Fujian Province of China(No.2021J05011)the Fundamental Research Funds for the Central Universities of Xiamen University(No.20720210033).
文摘In this work,we study the conic quadratic mixed-integer formulation for assortment optimization problem under the mixture of multinomial logit(MMNL)model.The MMNL model generalizes the widely studied multinomial logit choice model and can approximate any random utility model with an arbitrary additive error.An important operational decision problem in revenue management is assortment optimization problem,which aims to find a subset of products to make available to customers that maximizes the expected revenue of the retailer.It is known that assortment optimization problem under the MMNL model is NP-hard and inapproximable within any constant performance guarantee.Commonly used methods for solving such problem are heuristical approaches or customized combinatorial optimization approaches.In the meanwhile,studies related to global optimization approaches are relatively scarce.We propose an enhanced conic quadratic mixed-integer formulation for solving assortment optimization problem under the MMNL model with a higher computational efficiency.Furthermore,we conduct extensive numerical experiments to demonstrate that the proposed reformulation significantly outperforms the existing conic reformulations for assortment optimization under the MMNL model.
