最折衷權重分析應用於並列式程序資源分攤與產出分享之群體

Abstract

私營或公營企業的各單位，均需例行性地評比其下屬單位的績效。本研究針對組織中的一個特殊流程發展評量模式。組織中的每個受評單位(Unit of assessment, UOA)具有兩個子單位，此兩個子單位屬同一階層，每個程序可將投入項轉化為產出項，而資源可由子單位間分攤，產出也由子單位間分享，分攤與分享的比例為受上下限影響的決策變數，由系統來決定其值。本研究利用最折衷權重分析(The Most Compromised Weights Analysis, MCWA)決定組織的投入項共同權重與產出項共同權重，使組織的績效最大化，而每個UOA也可以得到其在總流程及兩個子單位的績效值。針對分析的結果，組織的管理階層可對各單位的投入項與產出項進行調整，以改善組織整體的績效。

This research introduces a procedure to solve a specific dynamic network DEA problem. A decision-maker governs a set of sub-units that convert three sets of inputs, xf, xh and xs, to three sets of outputs, yf, yh and ys by a process that contains two parallel sub-processes, f and h. xf and xh are dedicated to f and h respectively to produce yf and yh. f and h share xs and ys with given upper and lower bounds. We employ the linear programming model of the Most Compromised Weights Analysis to determine the common set of weights for the inputs and outputs so that the decision-maker’s performance is maximized. The proportions of each shared input and output assigned to f and h are also obtained. Base on the solutions, we compute the efficiency scores for each sub-unit in the overall process and sub-processes f and h. The decision-maker would enable to reallocate inputs and outputs to improve his/her performance. A illustrative numerical example is presented. The general model that multiple parallel sub-processes is introduced as well.