目標：針對緊急醫療救護系統之資源配置問題，研擬救護車數量與區位配置方案，以改善目前系統之績效。方法：收集臺南市93年1月至10月計13,871筆資料，分析並建立模擬模式。依各時段的案件空間分佈，利用空間分析的K-Means法，調整救護車之配置方式後，對各方案進行模擬分析，以評估其反應時間及救護車利用率的表現。結果：經模擬分析，整體反應時間以方案1 (14車) 表現最佳；而反應時間合乎標準的方案中，利用率以方案2 (13車) 最高；經綜合評估兩項指標，本研究認為方案6 (依時段調整車數) 為較佳之方案。結論：反應時間良寙與利用率大小，存有交互損益的關係，必須在兩者間取得平衡；本研究證實依案件之數量及空間分佈，機動調整救護車數量及駐點，除可改善緊急醫療救護系統反應時間的表現外，亦可提升利用率。

Objectives: The research attempts to assess alternatives of the number and locations of ambulances in order to improve emergency medical services systems performance.Methods: From January to October 2004 ambulances responded to 13,871 calls in Tainan City, Taiwan. According to the spatial distribution of calls of every shift, we used K-means, point pattern spatial analysis, to redeploy the ambulances. A simulation model based on the empirical data was developed to analyze response times and utilization ratios to evaluate the performance of each alternative.Results: Based on the simulation analysis, alternative #1 (14 ambulances) performs best on overall response time, while alternative #2 has the highestutilization ratio among those alternatives whose response times are acceptable. In comprehensive evaluation of the two indicators, alternative #6 (number of ambulances that varies with shifts) is superior.Conclusions: There is a trade-off relation between response time and utilization ratio. Our research shows that redeploying ambulances according to arrival frequency and spatial distribution of calls can improve not only the performance of their response time but also their utilization ratio.