2008-12-012024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/693608摘要：生技製藥業研發資源規劃之隨機動態規劃與實質選擇權模型 製藥業的新藥研發過程風險極高、耗時甚久，且需要大量的人力與資金投入。一般而言，即便是已經著手進行臨床實驗的化合物，其中大約只有百分之二十有潛力成為成功上市的藥物。而這些成功上市的藥物中，也只有五分之一能夠獲利。為了提昇新藥研發的成功率與期望報酬率，研發專案中資源的分配與規劃，是非常重要的課題。 本研究計畫將運用作業研究學領域中的隨機動態規劃與非線性最適化等數學規劃理論，針對新藥研發過程中，從化合物篩選至臨床實驗之前的研發過程，建立一整合性隨機模型。此隨機模型採用獲利能力指標以及內部報酬率兩種廣為採用的評估指標，並將進一步採納當今研發管理研究之實質選擇權分析法（Real Options Analysis），求解出新藥研發過程各階段的最佳資源規劃模式。本研究計畫將針對此隨機模型採用Visual C++ 撰寫專用之決策輔助軟體，以期能迅速有效地求算出最佳解，並進行資源規劃之敏感度分析與情境分析。 此隨機模型之目標方程式與限制式組，均為非線性且不連續函數，目前文獻中已知的動態規劃、隨機規劃、整數規劃、或是非線性規劃的求解演算法，皆無法直接套用在此隨機模型中求算最佳解。因此，本研究計畫將深入探討近似法或是數值分析之相關演算法技術，以求算此隨機模型之最佳解。 <br> Abstract: Stochastic Dynamic Programming and Real Options Models for Research Resource Planning in the Pharmaceutical Industry The pharmaceutical industry is highly competitive, and the discovery and development of new drugs is extremely expensive and time consuming. This research project aims to make a contribution to the task of improving the efficiency of pre-clinical research by building stochastic models with OR techniques such as stochastic dynamic programming, nonlinear optimization, and Real Options Analysis (ROA). The models to be developed in this research project investigate the numbers of scientists should be allocated to the successive, and where appropriate repeated, stages of a pre-clinical new drug discovery project so as to increase its profitability. The number of distinct series of compounds which should be explored in the search for candidate drugs is also investigated. Two widely-applied profitability criteria, Profitability Index and Internal Rate of Return, are considered, and the computer software designed to implement the optimization calculations and to carry out the sensitivity analysis is to be developed with Visual C++. The objective function and constraints incorporated in the models to be developed in this research project are all nonlinear or even discrete functions. Algorithms that have been proposed in the literature for dynamic programming, stochastic programming, integer programming, or nonlinear programming are not able to provide an efficient way to find the optimal solutions for the models to be developed in this research project. Thus methods regarding approximation or numerical analysis will be investigated.研發資源規劃實質選擇權研發管理決策分析Resource AllocationReal Options AnalysisOR in R&DDecision Analysis