In hotel management, traditional manual scheduling relies on experiential decision-making. It struggles to address dynamic challenges from occupancy fluctuations, periodic peaks, and sudden demands. This leads to issues like redundant labor and service gaps. To solve this problem, this study targets mid-range hotel management and proposes a staff scheduling algorithm integrating dynamic demand forecasting and multi-objective optimization. First, a demand forecasting model based on Long Short-Term Memory (LSTM) is built to predict periodic labor demands of each department. Second, a multi-constraint mathematical optimization model is established. Its objectives are minimizing total operating costs, minimizing service quality losses, and maximizing employee satisfaction. The model also considers skill matching, working hour compliance, and shift fairness. Finally, an improved Genetic Simulated Annealing Algorithm (GASA) is designed to solve the model. Through adaptive coding, a multi-objective fitness function, and a dynamic adjustment mechanism, the optimal scheduling scheme is generated. Empirical analysis shows the following. Under high occupancy (90%), the algorithm reduces daily operating costs by 15.3%, shortens average customer waiting time by 58.7%, decreases average employee overtime by 65.7%, and lowers customer complaint rate by 78.3%. Under sudden demand scenarios, the algorithm’s response time is less than 5 min, and the adjustment accuracy of scheduling schemes reaches 92%. This study provides a quantitative tool for dynamic staff scheduling in hotel management, offers scenario-specific references for multi-objective resource optimization in economy hotels with similar scale.
坊野恵子 et al. (Tue,) studied this question.