Search

Traffic flow prediction, a critical aspect of intelligent transportation systems, has been increasingly popular in the field of artificial intelligence, driven by the availability of extensive traffic data. The current challenges of traffic flow prediction lie in integrating diverse factors while balancing the trade-off between computational complexity and the precision necessary for effective long-range and large-scale predictions. To address these challenges, we introduce a Spatial-Temporal Selective State Space (ST-Mamba) model, which is the first to leverage the power of spatial-temporal learning in traffic flow prediction without using graph modeling. The ST-Mamba model can effectively capture the long-range dependency for traffic flow data, thereby avoiding the issue of over-smoothing. The proposed ST-Mamba model incorporates an effective Spatial-Temporal Mixer (ST-Mixer) to seamlessly integrate spatial and temporal data processing into a unified framework and employs a Spatial-Temporal Selective State Space (ST-SSM) block to improve computational efficiency. The proposed ST-Mamba model, specifically designed for spatial-temporal data, simplifies processing procedure and enhances generalization capabilities, thereby significantly improving the accuracy of long-range traffic flow prediction. Compared to the previous state-of-the-art (SOTA) model, the proposed ST-Mamba model achieves a 61.11\% improvement in computational speed and increases prediction accuracy by 0.67\%. Extensive experiments with real-world traffic datasets demonstrate that the \textsf{ST-Mamba} model sets a new benchmark in traffic flow prediction, achieving SOTA performance in computational efficiency for both long- and short-range predictions and significantly improving the overall efficiency and effectiveness of traffic management., Comment: 25 pages, 6 figures

Mobility-as-a-Service (MaaS) is an emerging business model driven by the concept of "Everything-as-a-Service" and enabled through mobile internet technologies. In the context of economic deregulation, a MaaS system consists of a typical two-sided market, where travelers and transportation service providers (TSPs) are two groups of agents interacting with each other through a MaaS platform. In this study, we propose a modeling and optimization framework for the regulation of two-sided MaaS markets. We consider a name-your-own-price (NYOP)-auction mechanism where travelers submit purchase-bids to accommodate their travel demand via MaaS platform, and TSPs submit sell-bids to supply mobility resources for the MaaS platform in exchange for payments. We cast this problem as a single-leader multi-follower game (SLMFG) where the leader is the MaaS regulator and two groups of follower problems represent the travelers and the TSPs. The MaaS regulator aims to maximize its profits by optimizing operations. In response to the MaaS regulator's decisions, travelers (resp. TSPs) adjust their participation level in the MaaS platform to minimize their travel costs (resp. maximize their profits). We analyze cross-group network effects in the MaaS market, and formulate SLMFGs without and with network effects leading to mixed-integer linear bilevel programming and mixed-integer quadratic bilevel programming problems, respectively. We propose customized branch-and-bound algorithms based on strong duality reformulations to solve these SLMFGs. Extensive numerical experiments conducted on large scale simulation instances generated from realistic mobility data highlight that the performance of the proposed algorithms is significantly superior to a benchmarking approach, and provide meaningful managerial insights for the regulation of two-sided MaaS markets in practice.

In the context of `Everything-as-a-Service', the transportation sector has been evolving towards user-centric business models in which customized services and mode-agnostic mobility resources are priced in a unified framework. Yet, in the vast majority of studies on Mobility as a Service (MaaS) systems, mobility resource pricing is based on segmented travel modes, e.g. private vehicle, public transit and shared mobility services. This study attempts to address this research gap by introducing innovative auction-based online MaaS mechanisms where users can bid for any amount of mode-agnostic mobility resources based on their willingness to pay and preferences. We take the perspective of a MaaS regulator which aims to maximize social welfare by allocating mobility resources to users. We propose two mechanisms which allow users to either pay for the immediate use of mobility service (pay-as-you-go), or to subscribe to mobility service packages (pay-as-a-package). We cast the proposed auction-based mechanisms as online resource allocation problems where users compete for MaaS resources and bid for travel time per trip. We propose (integer-) linear programming formulations to accommodate user bids based on available mobility resources in an online optimization approach. We show that the proposed MaaS mechanisms are incentive-compatible, develop customized online algorithms and derive performance bounds based on competitive analysis. Extensive numerical simulations are conducted on large scale instances generated from realistic mobility data, which highlight the benefits of the proposed MaaS mechanisms and the effectiveness of the proposed online optimization approaches.