Ignorant Critical Path Analysis with Triple Objectives for Event Networks and Its Application in Marine Ranching Emergency Decision

Emergency decision-making in event networks is often constrained by scarce historical data, conflicting expert judgments, and uncertain activity durations and costs. To address this problem, this paper develops an ignorance-event-network optimization framework that integrates GC-rule-based expert evidence fusion with a time-cost-risk multi-objective scheduling model. The GC rule is used to transform expert interval judgments into fused activity parameters and residual uncertainty-risk indicators. On this basis, a competitive multi-objective particle swarm optimization algorithm (CMOPSO) is applied to generate a Pareto set of feasible schedules, and TOPSIS is used to select a representative compromise solution. The framework is validated using a marine ranching ecological-security emergency decision network containing 62 real activities and two virtual start/finish nodes. Compared with the CPM baseline, the selected CMOPSO-TOPSIS schedule reduces project duration from 190.00 days to 156.41 days and slightly reduces the uncertainty-risk objective from 59.283 to 58.791, with total cost increasing from 326.80 to 361.46. The results show that the proposed framework can support transparent trade-off analysis among response duration, cost, and uncertainty risk in data-scarce marine ranching emergencies.