本研究摒棄傳統上假設金融資產的報酬率呈現常態分配或t分配，而以定態拔靴法的方式來模擬真實報酬率的分配，配合最小要求報酬限制模型，在既定風險下追求校務基金的報酬極大。本研究使用由下而上的資產配置方式，透過某些篩選機制挑選出優異的十檔股票、四檔債券型基金、三檔公債和定存形成投資組合，並以1998年3月到2006年2月共96筆月資料進行分析，決定最適資產配置。此外為檢驗最適資產配置的結果和績效，分別以不同的最小要求報酬 （0、-5％、-10％）、不同的時間長度（月、季、半年、一年）為調整期間以及考慮實際交易費用與否來進行樣本外測試，做法為從1998年3月開始，每次皆以五年共60個月的資料來對下一年（12個月）資產配置的報酬進行測試，以rolling的方式每次去掉第一個月（一季、半年、一年），再補上新一個月（一季、半年、一年）來測試最佳的資產配置方式。實證結果顯示在考慮交易成本之後，最小要求報酬為0（即不損失）的投資組合其Sharpe ratio皆為負數，代表投資組合的績效不如定存，還不如將錢都放在定存。因此本文將建議使用 為-5％或-10％資產配置方式，同時一個月調整一次投資組合，將可獲得較佳的投資績效。 / In this study, we implement the simulation of the real distributions of financial assets by means of the stationary bootstrap method instead of assuming normal distribution or t distribution. With the assistance of minimum required return model, we pursue the maximum profit under finite risk. We use the bottom-up asset allocation and select excellent investments by some criteria to form the portfolio, including four bond funds, ten stocks, three bonds and a time deposit. We use 96 monthly data from March 1998 to February 2006 to decide the best way for asset allocation . Besides, to make sure the asset allocation is practical, we also take transaction costs into account and conduct an out-of-sample test with different minimum required returns (0, -5%, -10%) and different holding periods (a month, a quarter, half a year, a year) to decide the best way for asset allocation. Starting from March 1998, we conduct an out-of-sample test with a solid 60-month data each time to test the return of the following year under specified asset allocation decisions. This is then done repeatedly by using the method of rolling, replacing the 1st-month data (1st-quarter data, 1st-half year data, 1st-year data) with new monthly data (quarterly data, semi-annual data, annual data) to find the best asset allocation. The Empirical result shows that after transaction costs are taken into account, the Sharpe ratios of the portfolio with the equal to zero are negative and the return are worse than the interest rate of the time deposit. Therefore, the asset allocation with equal to -5％ or -10％ will be recommended. Besides, monthly portfolio adjustment is better.