Abstract
Aiming at the vulnerability of nodes within collaborative task execution links and the slow response of link reconfiguration under limited communication conditions and sudden equipment failure scenarios at sea, and noting that existing approaches mainly rely on static optimization or offline planning without an integrated online reconfiguration framework, this paper proposes an A-star enhanced Multi-Objective Collaborative Reconfiguration (A-MOCR). By constructing a node state probability model and a multi-layer complex network framework, the proposed method sets maximization of task completion probability and network resilience as dual optimization objectives, and leverages the NSGA-III algorithm to generate primary links and corresponding backup link pools. Link reconfiguration is activated through real-time monitoring of node anomaly indicators; available surviving routes from the backup pool are prioritized for matching. If matching cannot be achieved, local A-star search or global path replanning will be initiated. Comparative simulation experiments against two baseline approaches demonstrate that the proposed A-MOCR achieves comparable mission success rates and link survival duration while significantly reducing the average reconfiguration time by over 50%. With equivalent task success probability, A-MOCR completes link reconfiguration within shorter time and maintains steady performance advantages under varying communication coverage parameters. Unlike conventional methods that require global re-planning after each failure, A-MOCR decouples offline optimization from online matching, enabling millisecond-level switching. The method realizes millisecond-level link switching and provides effective technical support for the adaptive reconfiguration of collaborative task execution links for marine multi-platform systems.