A strategic offshore investment portfolio model considering political risk and technology evolution in the semiconductor manufacturing industry
针对半导体制造业中技术快速进步与地缘政治紧张带来的不确定性，提出一个技术采纳组合模型，通过马尔可夫决策过程和差分进化算法优化长期投资利润，为战略离岸投资提供决策支持。
In the semiconductor manufacturing industry, decisions on strategic offshore investment and technology portfolio adoption under uncertainty are challenging. The rapid advancement of chip manufacturing technology and the political tensions among competitive countries intensify decision complexity. This study introduces a technology adoption portfolio model for strategic offshore investments, accounting for both technological and geopolitical uncertainties. The model aims to optimise the semiconductor technology adoption portfolio and the associated resources to maximise the expected profit in long-term investment. A Markov decision process (MDP) is proposed to model the problem, with each action presented as a mixed integer programming (MIP). A differential evolution algorithm (DE) is applied to find the optimal solution for each action. The problem is solved sequentially in two stages: (1) finding the optimal solution by DE for the MIP problem for each action,(2) solving the MDP problem using the backward recursive function. Due to the problem's computational complexity, a parallel-computing approach is adopted. A sensitivity analysis of the transition of technology adoption and geopolitical tension level is conducted to evaluate the tendency of investment in terms of expected profit. Marginal insights offer a robust decision-supporting strategy for expanding strategic offshore investment in semiconductor manufacturing within the context of geopolitical tensions.