運用核心密度估計最佳化緊急醫療服務資源的配置

Abstract

本研究旨在通過使用核密度估計（Kernel Density Estimation, KDE）來優化緊急服務位置的覆蓋率並縮短響應時間。我們採用 KDE 方法預測緊急事件的高發區域，並將這些區域作為最大覆蓋位置問題模型（Maximal Covering Location Problem, MCLP）的輸入，以改善資源分配。實驗結果顯示，KDE-MCLP 方法在平均響應時間和十分鐘內的覆蓋率方面顯著優於傳統的 MCLP 方法。具體而言，KDE-MCLP 方法的響應時間更短，且十分鐘內的覆蓋率更高，突顯了其在快速響應和廣泛覆蓋方面的顯著改進。這些結果表明，將 KDE 與 MCLP 結合使用可以顯著提升緊急資源配置的效率和效果，從而提高緊急醫療服務的整體效能。

This study aims to optimize the coverage and reduce the response time of emergency service locations using Kernel Density Estimation (KDE). We employed KDE to predict high-probability areas for emergency incidents and used these areas as inputs for the Maximal Covering Location Problem (MCLP) model to enhance resource allocation. The experimental results indicate that the KDE-MCLP method significantly outperforms the traditional MCLP method in terms of average response time and ten-minute coverage rate. Specifically, the KDE-MCLP method consistently shows shorter response times and higher ten-minute coverage rates, highlighting its significant improvements in rapid response and extensive coverage. These findings suggest that integrating KDE with MCLP can significantly enhance the efficiency and effectiveness of emergency resource allocation, ultimately improving the overall performance of emergency medical services.