Game-Theoretic and Algorithmic Analyses of Multi-Agent Routing under Crossing Costs

Coordinating the movement of multiple autonomous agents over a shared network is a fundamental challenge in algorithmic robotics, intelligent transportation, and distributed systems. The dominant approach, Multi-Agent Path Finding, relies on centralized control and synchronous collision avoidance, which often requires strict synchronization and guarantees of globally conflict-free execution. This paper introduces the Multi-Agent Routing under Crossing Cost model on mixed graphs, a novel framework tailored to asynchronous settings. In our model, instead of treating conflicts as hard constraints, each agent is assigned a path, and the system is evaluated through a cost function that measures potential head-on encounters. This ``crossing cost’’, which is defined as the product of the numbers of agents traversing an edge in opposite directions, quantifies the risk of congestion and delay in decentralized execution. Our contributions are both game-theoretic and algorithmic. We model the setting as a congestion game with a non-standard cost function, prove the existence of pure Nash equilibria, and analyze the dynamics leading to them. Equilibria can be found in polynomial time under mild conditions, while the general case is PLS-complete. From an optimization perspective, minimizing the total crossing cost is NP-hard, as the problem generalizes Steiner Orientation. To address this hardness barrier, we design a suite of parameterized algorithms for minimizing crossing cost, with parameters including the number of arcs, edges, agents, and structural graph measures. These yield XP or FPT results depending on the parameter, offering algorithmic strategies for structurally restricted instances. Our framework provides a new theoretical foundation for decentralized multi-agent routing, bridging equilibrium analysis and parameterized complexity to support scalable and risk-aware coordination.

💡 Research Summary

The paper introduces a novel routing model for multiple autonomous agents operating on a shared mixed graph, called Crossing‑Cost Multi‑Agent Routing (CC‑MAR). Unlike the classic Multi‑Agent Path Finding (MAPF) framework, which treats collisions as hard constraints and requires synchronized, centrally‑controlled execution, CC‑MAR assigns a soft “crossing cost” to each undirected edge: the product of the numbers of agents traversing that edge in opposite directions, multiplied by the edge’s weight. This cost captures the risk of head‑on encounters in asynchronous, decentralized settings where exact timing is unknown.

From a game‑theoretic perspective, the authors model the situation as a congestion game with a non‑standard cost function. They prove that the total crossing cost cost(P)=∑ₑ wₑ·x→ₑ·x←ₑ is exactly a Rosenthal potential function. Consequently, a pure‑strategy Nash equilibrium (NE) always exists. Moreover, best‑response dynamics converge in at most w_max·k·|E| steps, where w_max is the maximum edge weight, k the number of agents, and |E| the number of undirected edges. The equilibrium with minimum total cost coincides with the global optimum, implying a price of stability (PoS) of 1. However, the authors construct instances where the price of anarchy (PoA) is unbounded, showing that selfish behavior can lead to arbitrarily high total crossing cost.

Algorithmically, finding an NE is polynomial‑time solvable when edge weights are bounded by a polynomial in the input size (the dynamics can be simulated directly). In the general case with arbitrary integer weights, the problem becomes PLS‑complete, matching known hardness results for standard congestion games.

The optimization version—minimizing total crossing cost—generalizes the Steiner Orientation problem: deciding whether a zero‑cost solution exists is equivalent to orienting undirected edges so that each terminal pair (sᵢ,tᵢ) becomes strongly connected. Since Steiner Orientation is NP‑complete even on graphs of treewidth 2, CC‑MAR’s optimization is likewise NP‑complete.

To overcome this hardness, the authors turn to parameterized complexity. They present a comprehensive suite of algorithms:

• XP algorithm parameterized by the number of terminal pairs k – a dynamic‑programming approach that enumerates all possible orientations for the k paths. The authors argue that, because Steiner Orientation is W