A Dynamic Task Allocation Method for Heterogeneous UUVs in Communication-constrained Environments

This paper addresses the limitations of existing static Performance Impact (PI) algorithms in distributed task allocation for Unmanned Underwater Vehicle (UUV) clusters, which lack adaptability to the dynamic underwater environment with communication constraints. To overcome these challenges, we propose the Improved Dynamic Performance Impact (IDPI) algorithm, which extends the PI framework to dynamic task allocation. The IDPI periodically updates UUV positions and task completion status at discrete intervals, integrating time-triggered mechanisms with event-driven communication graph adjustments. This hybrid approach iterates between two phases: a task inclusion phase for candidate task selection and a consensus and task removal phase for conflict resolution. A key innovation of IDPI is the introduction of task-specific convergence flags, enabling UUVs to execute tasks immediately upon local consensus without waiting for global agreement - a departure from conventional PI algorithms that mandate full cluster-wide synchronization. Simulation results demonstrate that the proposed IDPI algorithm achieves rapid and efficient dynamic task allocation, significantly reducing decision latency while ensuring conflict-free assignments. The algorithm exhibits strong adaptability to dynamic environments with communication constraints, making it suitable for time-sensitive underwater missions requiring real-time coordination.