Scalable Algorithms with Provable Optimality Bounds for the Multiple Watchman Route Problem

A team of researchers including Srikar Gouru, Ariel Felner, and Jiaoyang Li has published a breakthrough paper on arXiv introducing scalable algorithms for the Multiple Watchman Route Problem (MWRP). The MWRP is a fundamental path-planning challenge where multiple agents (watchmen) must find routes that collectively ensure every point in an environment is visible to at least one agent. The team's key contribution is MWRP-CP3, an efficient optimal planner that combines novel state-space pruning methods with improved heuristic calculations. Their techniques identify and eliminate map areas guaranteed to be seen en route to other targets, dramatically shrinking the problem space.

The results are substantial: MWRP-CP3 reduces the required search space by more than 95% and executes over 200 times faster than previous optimal algorithms on standard 2D grid maps. For even larger-scale problems, the researchers developed suboptimal algorithms with provable quality bounds, including MxWA* (a weighted A* variant for makespan problems) and anytime variations. These suboptimal solvers can handle maps three times larger than those solvable by the optimal MWRP-CP3, pushing the boundaries of practical applicability. The work includes methods to iteratively improve existing solutions by solving decomposed sub-problems.

The research provides both theoretical guarantees and practical performance, with the complete codebase and video demonstrations made publicly available. This advancement directly addresses scalability limitations that have hindered real-world deployment of multi-agent surveillance and coverage systems, offering new tools for robotics, automated inspection, and security logistics where efficient team coordination is critical.