Douglas G. Macharet

João Vitor de Carvalho Silva, Douglas G. Macharet

The ability to coordinate actions across multiple agents is critical for solving complex, real-world problems. Large Language Models (LLMs) have shown strong capabilities in communication, planning, and reasoning, raising the question of whether they can also support effective collaboration in multi-agent settings. In this work, we investigate the use of LLM agents to solve a structured victim rescue task that requires division of labor, prioritization, and cooperative planning. Agents operate in a fully known graph-based environment and must allocate resources to victims with varying needs and urgency levels. We systematically evaluate their performance using a suite of coordination-sensitive metrics, including task success rate, redundant actions, room conflicts, and urgency-weighted efficiency. This study offers new insights into the strengths and failure modes of LLMs in physically grounded multi-agent collaboration tasks, contributing to future benchmarks and architectural improvements.

Luciano E. Almeida, Douglas G. Macharet

This paper introduces an extension to the Orienteering Problem (OP), called Clustered Orienteering Problem with Subgroups (COPS). In this variant, nodes are arranged into subgroups, and the subgroups are organized into clusters. A reward is associated with each subgroup and is gained only if all of its nodes are visited; however, at most one subgroup can be visited per cluster. The objective is to maximize the total collected reward while attaining a travel budget. We show that our new formulation has the ability to model and solve two previous well-known variants, the Clustered Orienteering Problem (COP) and the Set Orienteering Problem (SOP), in addition to other scenarios introduced here. An Integer Linear Programming (ILP) formulation and a Tabu Search-based heuristic are proposed to solve the problem. Experimental results indicate that the ILP method can yield optimal solutions at the cost of time, whereas the metaheuristic produces comparable solutions within a more reasonable computational cost.

Armando Alves Neto, Víctor C. da Silva Campos, Douglas G. Macharet

A critical metric of the coverage quality in Wireless Sensor Networks (WSNs) is the Minimal Exposure Path (MEP), a path through the environment that least exposes an intruder to the sensor detecting nodes. Many approaches have been proposed in the last decades to solve this optimization problem, ranging from classic (grid-based and Voronoi-based) planners to genetic meta-heuristics. However, most of them are limited to specific sensing models and obstacle-free spaces. Still, none of them guarantee an optimal solution, and the state-of-the-art is expensive in terms of run-time. Therefore, in this paper, we propose a novel method that models the MEP as an Optimal Control problem and solves it by using a Semi-Lagrangian approach. This framework is shown to converge to the optimal MEP while also incorporates different homogeneous and heterogeneous sensor models and geometric constraints (obstacles). Experiments show that our method dominates the state-of-the-art, improving the results by approximately 10% with a relatively lower execution time.

Douglas G. Macharet, Armando Alves Neto, Daigo Shishika

Different applications, such as environmental monitoring and military operations, demand the observation of predefined target locations, and an autonomous mobile robot can assist in these tasks. In this context, the Orienteering Problem (OP) is a well-known routing problem, in which the goal is to maximize the objective function by visiting the most rewarding locations, however, respecting a limited travel budget (e.g., length, time, energy). However, traditional formulations for routing problems generally neglect some environment peculiarities, such as obstacles or threatening zones. In this paper, we tackle the OP considering Dubins vehicles in the presence of a known deployed sensor field. We propose a novel multi-objective formulation called Minimal Exposure Dubins Orienteering Problem (MEDOP), whose main objectives are: (i) maximize the collected reward, and (ii) minimize the exposure of the agent, i.e., the probability of being detected. The solution is based on an evolutionary algorithm that iteratively varies the subset and sequence of locations to be visited, the orientations on each location, and the turning radius used to determine the paths. Results show that our approach can efficiently find a diverse set of solutions that simultaneously optimize both objectives.