LLM-Based Multi-Agent Collaboration for Constrained Multi-Objective Container Placement

XAI-Guided Conservative Decentralized Execution for Offline Multi-Agent Network Slicing

X-CODE is an explainable offline MARL that operates offline without environmental interaction, nor inter-agent communication, nor inter-agent communication, and exploits explainability-aware reward shaping to modify the relative preference among joint offline transitions during centralized training to improve decentralized resource-allocation behavior.

ICBBA-ACO-Based Multi-Robot Task Allocation for Smart Charging Stations

Smart charging stations require mobile charging robots to respond to dynamically arriving charging requests with heterogeneous priorities, varying travel costs, and uneven workloads while maintaining online scheduling feasibility. Conventional single-layer approaches often optimize task assignment or route ordering separately, which limits their ability to coordinate allocation quality, route efficiency, and workload regulation under real-time constraints. This study proposes a hierarchical improved consensus-based bundle algorithm–ant colony optimization (ICBBA-ACO) framework for dynamic multi-robot task allocation. The upper ICBBA layer combines deterministic task clustering, intra-cluster greedy bundling, conflict resolution, and feedback-guided workload-aware reassignment, while the lower ACO layer refines the visiting order of unstarted tasks under fixed ownership using the same normalized four-objective scheduling cost. Complete decision time is evaluated separately against a 200ms online requirement, and estimated motion energy is retained only as a distance-derived auxiliary indicator. In a five-method comparison over 100 paired scenarios, ICBBA-ACO achieves a mean composite objective of J=0.663052, a mean decision time of 33.07ms, and 100% deadline compliance. GA-MRTA obtains a lower unconstrained mean objective of J=0.615790, but requires approximately 2199.30ms on average and satisfies the 200ms requirement in only 8.89% of the evaluated updates. Thus, ICBBA-ACO provides the lowest mean objective among the compared methods that maintain full deadline compliance, demonstrating a favorable quality–runtime trade-off within the tested operating range. ROS-based engineering verification further completes all 15 repeated trials and all 48 verification tasks with no recorded invariant violations.

The dynamic multi-agent approach to solving the traveling salesman problem

This paper investigates the multi-agent traveling salesman problem with a common depot and dynamic distribution of service points. Unlike traditional approaches that require preliminary clustering and rigid assignment of points to agents, the proposed method ensures adaptive formation of service zones directly during route construction. The distribution principle is based on minimizing the distance to the agent's initial position, which leads to the formation of spatially separated and compact clusters. Routing within the formed subsets is performed using a modified ant colony algorithm, in which the transition probability accounts for, in addition to pheromone trail intensity and heuristic information inversely proportional to distance, an additional correcting factor reflecting the candidate's proximity to the starting vertex. This architecture allows solving clustering and routing problems in a coordinated manner within a single optimization process, eliminating the need for a preliminary partitioning stage. To reduce computational complexity, multi-threaded processing of formed clusters is implemented, ensuring algorithm scalability as the number of agents increases. Results of computational experiments on standard test sets of various dimensions confirm the effectiveness of the proposed approach in terms of minimizing the total route length. It has been established that integrating the proximity factor to the initial position into the heuristic function contributes to the formation of more compact trajectories compared to the basic version of the ant colony algorithm. With comparable computational costs, the dynamic distribution scheme provides an improvement in the objective function by an average of 6% and also enhances the system's adaptability to dynamic changes in input parameters.

MOAE: Multi-Objective Agent Evolution with Pareto-Preserving Search

As LLM-based agents continue to advance, their evaluation has become increasingly multifaceted: a capable agent must not only achieve high task completion accuracy but also perform well in interaction quality, safety, and efficiency, raising a central question: can these objectives be optimized simultaneously? Existing methods have considered multiple objectives, but many collapse heterogeneous measurements into a fixed scalar score. Such scalarization depends on metric normalization and preference weights and may discard candidates that represent useful deployment trade-offs. We introduce Multi-Objective Agent Evolution (MOAE), which organizes iterative in-context refinement as a Pareto-preserving evolutionary search over complete agent rollouts. Given a limited rollout budget, MOAE maintains an empirical archive of non-dominated candidates, uses objective-specific diagnostics to guide offspring generation, and applies constraint-aware selection only at deployment. This separates candidate preservation during search from the preference used to return a final solution. The procedure requires no parameter updates and allows each objective to be replaced by any measurable property, which we instantiate as task performance, trajectory quality, and safety. Experiments on TravelPlanner and AgentDojo show that MOAE consistently improves task performance and trajectory quality while maintaining strong safety under matched rollout budgets. Search-behavior analysis further shows that Pareto preservation expands the attainable objective region and increases the frequency of joint improvement. These results demonstrate the potential of Pareto-preserving in-context evolution for optimizing multiple agent properties without committing to a fixed scalarization during search.

Multi-Agent Scheduling for Large-Scale On-Road Testing in Intelligent Transportation Systems

Uncertainty-aware multi-objective optimization for rebalancing bike shared systems

Overnight rebalancing in dock-based bike-sharing systems requires routing a limited fleet of trucks before user activity begins. This article formulates static rebalancing problem under stochastic demand uncertainty as a bi-objective combinatorial optimization problem that selects truck routes and visited stations. The first objective minimizes total travel distance. The second objective minimizes scenario-weighted unmet demand under a finite set of demand scenarios derived from historical station-status data. A deterministic recourse evaluation simulates truck loads and station inventories along each route and computes unmet demand for visited and unvisited stations. The article applies two multi-objective evolutionary algorithms, NSGA-II and MOEA/D, using a permutation–partition encoding and relocate-based operators that implement a 1–0 relocate neighborhood between routes. A roulette-wheel-based relocation operator (BB2) biases move selection by the induced change in route distance. Experiments on the Barcelona Bicing network with 518 stations and on clustered subinstances show that NSGA-II attains higher hypervolume and larger non-dominated sets, whereas MOEA/D attains lower runtime; an ablation analysis shows that BB2 improves coverage and proximity indicators.