Bin Xin

Accidental leakage of hazardous gases in indoor environments poses significant risks to human health and the environment, necessitating rapid and effective source-seeking. The main challenges of multi-robot indoor intermittent source seeking are twisted plume distributions caused by obstacles and the tendency of robots to cluster around outdated concentration peaks. To tackle these challenges, a dynamic mode-switching multi-robot search method is proposed, combining a probabilistic plume-tracking strategy and an information gain-based exploration strategy. In this method, the robots are dynamically assigned to either tracker or explorer mode based on their measurements. The tracker robots to evaluate candidate source locations and follow the plume toward the leak source, while explorer robots calculate information gain based on unexplored frontiers. Mode-switching between tracker and explorer facilitates an effective balance between plume tracking and environmental exploration. The method further incorporates Voronoi diagrams for dynamic partitioning and a time-decaying scaling factor to mitigate clustering around local or outdated peaks. Experimental results in simulated and real environments demonstrate the method’s superior performance.

This paper investigates a novel bi-objective moving-target traveling salesman problem with uncertainty, where the target trajectories are uncertain. The objective is to maximize the number of visited targets within a bounded region while minimizing the travel cost. This problem is formulated as a bi-objective two-stage stochastic program with recourse. The first-stage decision determines a planned visiting sequence before the realization of target trajectories, and the second-stage recourse decision allows the agent to return to the depot in advance when the remaining targets are expected to leave the bounded region. We adopt the sample average approximation approach to approximate the expected objective values by empirical average over a finite set of sampled scenarios. To efficiently approximate the Pareto front, we propose an Indicator-based Artificial Bee Colony (I-ABC) algorithm. We introduce a ratio-based binary indicator into both the environmental selection and evolutionary mechanisms, incorporate a parallel-distance measure to preserve population diversity, and adopt an elitist strategy in the scout bee phase to enhance convergence. Extensive experiments in simulated environments demonstrate that the proposed I-ABC algorithm outperforms the state-of-the-art algorithms.

This study develops an integrated optimization model for vehicle-drone collaborative delivery in low-altitude logistics, incorporating heterogeneous social vehicles via a dynamic quotation and negotiation process. A hybrid heuristic framework combining mixed-integer programming and event-driven simulation is proposed to handle dynamic vehicle availability and pricing uncertainty. The approach coordinates dynamic pricing, capacity-aware grouping, and cost-sensitive routing decisions, effectively balancing delivery efficiency with operational costs to maximize profit in time-sensitive last-mile distribution.