A joint response model for matched decision makers: exploring decision making mechanism for mutually-selected agents

Abstract

Among these methods, econometric modeling does not assume any behavioral rules and allows the collected data to elucidate this matter. However, existing econometric models are not able to behavioral-consistently capture these new phenomena, such as intricate matching network, mutual selection, and intensive joint decision making. Therefore, this dissertation develops an innovative econometric model to fill the void. Specifically, the proposed model consists of two parts: The first part explains the matching process in a many-to-many matching structure; The second part characterizes the joint decision making process of mutually-selected decision makers. The two parts are integrated by recognizing their dependency that is essentially a sample selection process: a joint response is only observed for matched decision makers.

Burgeoning information technology innovations and the wide adoption of GPS devices have greatly changed the transportation system. For travelers, the real-time ridesharing platforms (e.g., Uber) allow drivers and riders to interact, pair up, and jointly decide on departure time and routes. In freight transportation, the prosperity of e-commerce leads to individualized real-time seller-buyer matching and their joint decisions on delivery modes and time windows. Transportation agents mutually select, or get matched with their counter-partners, and jointly make decisions on a set of matters that can be measured as linear, ordinal, or categorical values. Popular and potential methodologies of understanding these emerging collaborative phenomena include agent-based modeling, cooperative game theory, optimization-based approaches, and econometric modeling.

The proposed model is estimated using a Bayesian Markov-Chain Monte-Carlo approach with data augmentation. The likelihood functions and posterior distributions are derived for the ordinal and multinomial joint response outcomes respectively. Then, a simulation dataset is generated based on pre-defined parameters, and parameter recovery capability is measured as an indicator of model performance. A series of simulation datasets are further generated with respect to different parameter settings to evaluate the sensitivity of parameter recovery capability. Lastly, two empirical transportation applications are presented to demonstrate applicable values of the proposed model. The first application investigates flight on-time performance considering the mutual selection and joint responses of airlines and airports. The second application analyzes freight carriers’ responses to hypothetical toll increases with the consideration of their interactions with freight customers.