An evolutionary competition-aware construction heuristic for a stochastic resource allocation problem

The stochastic resource allocation (SRA) problem involves determining optimal resource assignments when the effectiveness of resources in completing assigned tasks is probabilistic. Weapon-target allocation (WTA) is a widely studied form of SRA. Existing heuristic approaches for WTA often adopt search methods with cumbersome solution encodings or constructive methods with greedy criteria that overlook competition among resource-task allocation pairs (RTAPs), potentially limiting solution quality. To efficiently obtain high-quality WTA solutions, this paper proposes an evolutionary competition-aware construction heuristic, termed ECACH, whose core lies in constructing solutions through a competition-aware construction (CAC) mechanism. A parameterized fuzzy evaluation system is employed as a flexible construction criterion, with its parameters optimized by an evolutionary algorithm to generate diverse evaluation values for RTAPs. Based on these evaluations, CAC adjusts each candidate RTAP’s construction priority by considering competing RTAPs that share the same resource or task. This integration of fuzzy-system-based evaluation and competition-aware priority adjustment reduces the shortsightedness of conventional greedy construction methods and enables more promising RTAPs to be selected earlier, thereby yielding better convergence and solution quality. Experiments on WTA instances with varying scales and resource-to-task ratios demonstrate that ECACH achieves better solution quality and competitive runtime performance under the same computational environment compared with state-of-the-art WTA algorithms.