An Intelligent Decision Support System for Dynamic Resource Allocation Using Hybrid Rule-Based and Learning Models

Dynamic resource allocation remains a persistent challenge in intelligent systems, particularly in environments where system conditions change rapidly and decisions must balance efficiency, stability, and interpretability. Traditional rule-based approaches offer transparency but struggle to adapt under uncertainty, while learning-based methods adapt well yet often operate as opaque decision makers. This paper presents an intelligent decision support system that combines both approaches in a unified hybrid framework for dynamic resource allocation. The proposed system integrates a rule-based reasoning module with a lightweight learning model, enabling decisions that are both adaptive and explainable. Rules encode domain knowledge and operational constraints, while the learning component captures patterns from historical and real-time data to anticipate resource demand. A decision fusion mechanism is used to resolve conflicts between the two components and the consistent and reliable allocation results are achieved. The system is tested in a simulated environment that represents different situations of work load and limited resources. Measures used in the evaluation of performance are resource-utilization efficiency, accuracy of decisions and response time. It has been shown that the hybrid method has more stable allocation behaviour and greater responsiveness than do standalone rule-based and learning-based baselines. The results indicate that explicit knowledge together with data-driven learning is a reality that can lead to intelligent and reliable decision-making in resource-constrained systems. The work offers a systematic base of future extensions with adaptive rule evolution and real world deployment conditions.