Dynamic Allocation of Critical Chain Buffers for Multi-objective Optimization and Project Schedule Risk Management

Considering duration, cost, and quality constraints, this study develops a multi-objective optimization model for project scheduling. The framework employs trapezoidal fuzzy numbers to characterize activity durations and a bell-shaped function to define the duration-quality trade-off. An optimal schedule is derived using a genetic algorithm. To mitigate uncertainties in construction progress, four duration risk categories—resource, operational, human, and environmental—are quantified through risk coefficients and weighting factors to establish a robust buffer sizing mechanism. Upon 50% completion of the planned workload, a grey GM(1,1) model is implemented to forecast activity finish times based on real-time progress monitoring. Furthermore, a dynamic buffer allocation model is established by integrating initial buffers, early completion gains, and cumulative buffer consumption. Application to a pumped storage power station project demonstrates that the proposed method reduces the total duration by 18 days, confirming its efficiency and practical utility in complex engineering environments.