Online ride-hailing platforms aim to deliver efficient mobility-on-demand services, often facing challenges in balancing dynamic and spatially heterogeneous supply and demand. Existing methods typically fall into two categories: reinforcement learning (RL) approaches, which suffer from data inefficiency, oversimplified modeling of real-world dynamics, and difficulty enforcing operational constraints; or decomposed online optimization methods, which rely on manually designed high-level objectives that lack awareness of low-level routing dynamics. To address this issue, we propose a novel hybrid framework that integrates large language model (LLM) with mathematical optimization in a dynamic hierarchical system: (1) it is training-free, removing the need for large-scale interaction data as in RL, and (2) it leverages LLM to bridge cognitive limitations caused by problem decomposition by adaptively generating high-level objectives. Within this framework, LLM serves as a meta-optimizer, producing semantic heuristics that guide a low-level optimizer responsible for constraint enforcement and real-time decision execution. These heuristics are refined through a closed-loop evolutionary process, driven by harmony search, which iteratively adapts the LLM prompts based on feasibility and performance feedback from the optimization layer. Extensive experiments based on scenarios derived from both the New York and Chicago taxi datasets demonstrate the effectiveness of our approach, achieving an average improvement of 16% compared to state-of-the-art baselines.

In this paper, we proposed a conflict-free routing strategy combined with scheduling for Terminal Manoeuvring Area (TMA) multi-aircraft to guarantee a safe separation. By incorporating Standard Terminal Arrival Routes (STARs) as route constraints, a mixed-logic model is designed to maximize the runway throughput while ensuring minimum separation between aircraft and avoiding overtaking on each STARs segment. Control techniques such as speed recommendation and holding operations are employed to the model to address potential conflicts. Three different algorithms are developed to solve the model: branch and bound with mixed-integer linear programming, multi-agent pathfinding with constraint programming, and meta-heuristics with evolutionary neighborhood search. These algorithms are tested on multiple cases of varying scales. Finally, we demonstrate the advantages of the proposed three algorithms by simulating realistic scenarios and comparing the results with Singapore ADS-B (Automatic dependent Surveillance-Broadcast) historical dataset. In one hour testing, results show that our method could reduce the last aircraft landing time nearly 10 minutes and save more than 80 minutes for total flight travel times for all aircraft, as well as non-vectoring flight trajectories, which indicates its potential to be used as an auxiliary decision-making tool for Air Traffic Controllers (ATCOs).

Road accidents wreck lives. Could technology stop them from happening? Driving better road safety with technology and artificial intelligence are the key elements considered by several carmakers. The key aspect of transportation in the future is to build an ecosystem comprising autonomous, connected, electric and shared mobility. The evolution of autonomous vehicles (AVs) can potentially aid transportation to people and be deployed to resolve mobility-related pain for drivers and safety on roads while changing lanes. Thus, the intelligent assistance system should be smart enough to track such vehicles while deviating into another lane. In this paper, we propose a lane change prediction framework for feature learning, with the aim to have a deep and comprehensive understanding of lane change behaviors, meanwhile, reach a high performance based on the selected features. A time-step dataset with more than 1000 features is constructed from vehicle trajectory data. To identify the key features involved in the original feature set, an XGBoost-based three-step feature learning algorithm is proposed, which integrates the feature importance ranking, metric selection and recursive feature elimination. After analyzing the accuracy of test data from different time segment positions, the sliding window method is applied on a time-step dataset with filtered features to properly select time segments, which are flattened into corresponding time-series dataset for model prediction. In our case studies, a publicly available dataset, Next Generation SIMulation (NGSIM), is adopted to conduct experiments of feature learning and lane change prediction, where we achieved a new state-of-art accuracy with 97.6% under the time-series data of 75 selected features and 1-second window size with predictor XGBoost after adopting proposed three-step method, which is superior to the other state-of-the-art feature selection methods.

A multi-bus dispatching strategy is proposed for a ring-shaped road bus transport system, which allows dispatching single bus or multiple buses and incorporates volume dynamics on both buses and stations. Also, the passengers’ perceived waiting time is firstly formulated as one part of the cost function to take passengers’ anxiety into account, and thereby improving the bus quality of service of bus operations. At upstream stations, as many passengers as possible will board the bus, which leads to the less space remaining on the bus and thus the enlongated wait for passengers at downstream stations. With the aim to avoid such phenomenon, the bus boarding control is implemented in the passengers’ boarding process captured by a simultaneous loading model to provide boarding opportunities for the waiting passengers at downstream bus stations. The formulated problem is tackled in two different scenarios, i.e., either with a linear cost or with a nonlinear cost. The linear cost, incorporating the passengers’ actual waiting time and the bus utilization, is firstly converted into a Mixed Integer Linear Programming (MILP) problem, and is solved by the commercial solver Gurobi. With the computational complexity as a concern, two different evolutionary algorithms, Genetic Algorithm (GA) and Harmony Search algorithm (HS), are also adopted to solve the problem in real time. In Scenario 2, the nonlinear cost, integrating the passengers’ perceived waiting time and the bus utilization, is directly solved by both GA and HS. Finally, case studies are provided to illustrate the efficiency of our proposed strategy by comparing with the traditional bus schedule strategies, as well as analyzing the different impacts of the bus loading process when either passengers’ actual waiting time or passengers’ perceived waiting time are taken into account.

The continuing increase of the on-road private cars is contributing to a deterioration of the urban traffic system. Public transportation is widely used to tackle this issue due to its large ridership. In this paper, we propose a multi-bus dispatching strategy combined with the boarding and holding control (MBDBH) to improve bus utilization and further decrease the passenger excess delay. Dispatching adjustments and operation control are taken into account in the system. At the dispatching level, on the one hand, either a bus platoon or a single bus can be dispatched for each trip to provide adaptive bus capacity to match the highly-fluctuated stop demands, on the other hand, we adjust the bus dispatching time based on the existing timetable to minimize passenger excess waiting time to a large extent. Meanwhile, the operation level incorporates both holding strategy and boarding limit strategy to bring more flexible adjustments in improving bus service. Besides the efficiency, we also minimize the headway variation in order to maintain a high system reliability. The problem is formulated as a Mixed Integer Nonlinear Programming (MINP) problem, which is solved by the commercial solver Gurobi. With the computational complexity as a concern, we propose a distributed algorithm to implement dual decomposition based on the partial Lagrangian relaxation. Finally, numerical examples are investigated to illustrate the significant time reduction of distributed algorithm and the efficiency of our proposed strategy: The proposed MBDBH model can reduce roughly 50% and 30% of remaining passenger volumes when compared with the timetable-based fixed schedule and the optimized single-bus dispatching schedule, respectively.

In order to efficiently balance the current and future air traffic demands with the system capacity, a proper Air Traffic Flow Management (ATFM) approach is required. The current focus of ATFM is generally on optimally utilizing the available airspace and airport capacities, while maintaining the required safety separation between aircraft. Yet, only a minor focus is given to the inherent uncertainty in the Air Transportation System (ATS), especially to its adverse effect on safety and day-to-day operations. To this end, we propose an ATFM framework scrutinizing the stochastic nature of ATS through a chance-constraint-based probabilistic approach. Moreover, anticipating the high volumes in air traffic in the future, we propose to split the model into two stages, in which the first stage scrutinizes the behavior of a set of flights as a flow, while the second stage transforms them into individual flight plans, enhancing scalability. The two models are formulated as an Integer Linear Programming (ILP) problem, and a Mixed Integer Linear Programming (MILP) problem at stages I and II, respectively. The NP-hard nature of the overall problem is minimized by transforming the problem into a Maximum Weighted Independent Set (MWIS) finding problem.

Conflicts between pedestrians and vehicles are one of the common safety issues at signalized intersections. Pedestrian Flashing GREEN (FG), a time interval for pedestrians on crosswalks to safely finish crossing before the next phase occurs, may fail to clear the crosswalk in the allotted time, due to significant pedestrian non-compliant behavior. In this manner, probability of pedestrian-vehicle exposures increases when non-compatible vehicle flows are released at the next immediate phase. This paper seeks to address this issue by presenting a traffic signal control strategy for urban traffic networks that aims to minimize vehicle traveling delay (increase efficiency) as well as pedestrian crossing risk (increase safety). First, a macroscopic model for pedestrian-vehicle mixed-flow networks is proposed. Considering the high-incidence rate of pedestrian violations during FG, an additional Dynamic All RED (DAR) phase is introduced at the end of each FG period, whose duration is adaptively adjusted according to the number of non-compliant pedestrians. With computational complexity being a concern for our model, an evolutionary algorithm with repairing mechanism (EARM), is proposed to solve our problem. Case studies are provided to illustrate the potential impact of the pedestrian movement to the vehicle traffic networks when pedestrian safety is considered in the system, as well as the efficacy of our traffic light control strategy for pedestrians and vehicles on risk reduction.

This paper addresses a bi-objective urban traffic light scheduling problem (UTLSP), which requires minimizing both the total network-wise delay time of all vehicles and total delay time of all pedestrians within a given finite-time window. First, a centralized model is employed to describe the UTLSP, where the cost functions and constraints of the two objectives are presented. A non-domination strategy-based metric is used to compare and rank solutions based on the two objectives. Second, metaheuristics, such as harmony search (HS) and artificial bee colony (ABC), are implemented to solve the UTLSP. Based on the characteristics of the UTLSP, a local search operator is utilized to improve the search performance of the developed optimization algorithms. Finally, experiments are carried out based on the real traffic data in Jurong area of Singapore. The HS, ABC, and their variants with the local search operator are evaluated in 19 case studies with different scales and time windows. To the best of our knowledge, this paper is the first of its kind to solve bi-objective traffic light scheduling problems in the literature. To demonstrate the effectiveness of the proposed algorithms in dealing with bi-objective optimization in traffic light scheduling, they are compared to the classical non-dominated sorting genetic algorithm II (NSGAII) with and without the local search operation. The comparisons indicate that our algorithms outperform the NSGAII algorithm with and without the local search operator for solving the UTLSP.

This paper presents a macroscopic model for pedestrian-vehicle mixed-flow network and a traffic signal scheduling strategy for both pedestrians and vehicles. We first propose a novel mathematical model of pedestrians crossing a junction. By combining a link-based vehicle network model, a traffic light scheduling problem is formulated with the aim to strike a good balance between pedestrians' needs and vehicle drivers' needs. The problem is first converted into a mixed-integer linear programming (MILP) problem via a novel transformation procedure, which is solvable by several existing solvers, e.g., GUROBI. Then a meta-heuristic method called discrete harmony search (DHS) algorithm is also adopted to reduce the computational complexity in MILP. Numerical simulation results are provided to illustrate the effectiveness of our real-time traffic light scheduling strategy for pedestrians and vehicles, and the potential impact of the pedestrian movement to the vehicle traffic flows.

Efficient flight routing and scheduling play an important role in air traffic flow management, which aims to maximize the utilization of airport and enroute capacities to ensure safety and efficiency of air transportation. In this paper, we first propose a novel discrete-time flow dynamic model for an air traffic network, consisting of airports, waypoints, and air links, upon which we formulate an air flow routing and scheduling problem as an integer linear programming problem. Considering the NP-hard nature of the problem, we present a novel hierarchical flow routing and scheduling approach, where the hierarchical architecture is derived naturally from the network containment relationship, and computation is carried out in a bottom-up manner, which relies on an incremental strategy. On the resulting flow routes and schedules, a heuristic algorithm is carried out to determine flight plans for individual aircrafts. The effectiveness of the proposed hierarchical approach is illustrated by air traffic data in four flight information regions in the association of Southeast Asian nations.