A Hidden Markov Model Based Extended Case-Based Reasoning Algorithm for Relief Materials Demand Forecasting
نویسندگان
1 Department of Industrial Management, Faculty of Management, University of Tehran, I. R. Iran
2 Department of Industrial Management, Faculty of Management, University of Tehran, I. R. Iran
3 Department of Industrial Management, Faculty of Management, University of Tehran, I. R. Iran
4 Department of Industrial Management, Faculty of Management, University of Tehran, I. R. Iran
doi
10.22052/mir.2023.253148.1420چکیده
In emergency situations, accurate demand forecasting for relief materials such as food, water, and medicine is crucial for effective disaster response. This research is presented a novel algorithm to demand forecasting for relief materials using extended Case-Based Reasoning (CBR) with the best-worst method (BWM) and Hidden Markov Models (HMMs). The proposed algorithm involves training an HMM on historical data to obtain a set of state sequences representing the temporal fluctuations in demand for different relief materials. When a new disaster occurs, the algorithm first determines the current state sequence using the available data and searches the case library for past disasters with similar state sequences. The effectiveness of the proposed algorithm is demonstrated through experiments on real-world disaster data of Iran. Based on the results, the forecasting error index for four relief materials is less than 10\%; therefore, the proposed CBR-BWM-HMM is a strong and robust algorithm.
