TY - JOUR
T1 - Predicting the competitive relationships of industrial production between Taiwan and China using Lotka–Volterra model
AU - Tsai, Bi-Huei
PY - 2017/5/28
Y1 - 2017/5/28
N2 - This work is the first to apply Lotka–Volterra model combined with genetic algorithm (GA) to predict the production relationships of high-tech industry among different areas. Previous studies analysed the trade interdependency among various countries, but few studies have highlighted the quantitative evidence of production relationships. Thus, this study utilizes motherboard shipment volumes to predict the competitive relationships of industrial production on both sides of the Taiwan Strait. Specifically, this work uses simultaneous non-linear least square regression in combination with GAs for numerical parameter optimization of the proposed Lotka–Volterra model. The results of parameter estimation reveal that shipment growth in China substantially promotes that in Taiwan, whereas the shipment growth in Taiwan curtails that in China. The standard deviation of the estimated parameters from the 3000 iterated simulations is small, confirming the reliability and stability of our parameter estimations. According to equilibrium analysis, the results of Lyapunov function prove that the shipments of China and Taiwan will reach a stable long-term equilibrium. The potential production from China will ultimately be nearly 16 times as large as that from Taiwan. Finally, the analytical results of forecast accuracy confirm that Lotka–Volterra model performs better than conventional S-curve diffusion model in predicting motherboard shipments.
AB - This work is the first to apply Lotka–Volterra model combined with genetic algorithm (GA) to predict the production relationships of high-tech industry among different areas. Previous studies analysed the trade interdependency among various countries, but few studies have highlighted the quantitative evidence of production relationships. Thus, this study utilizes motherboard shipment volumes to predict the competitive relationships of industrial production on both sides of the Taiwan Strait. Specifically, this work uses simultaneous non-linear least square regression in combination with GAs for numerical parameter optimization of the proposed Lotka–Volterra model. The results of parameter estimation reveal that shipment growth in China substantially promotes that in Taiwan, whereas the shipment growth in Taiwan curtails that in China. The standard deviation of the estimated parameters from the 3000 iterated simulations is small, confirming the reliability and stability of our parameter estimations. According to equilibrium analysis, the results of Lyapunov function prove that the shipments of China and Taiwan will reach a stable long-term equilibrium. The potential production from China will ultimately be nearly 16 times as large as that from Taiwan. Finally, the analytical results of forecast accuracy confirm that Lotka–Volterra model performs better than conventional S-curve diffusion model in predicting motherboard shipments.
KW - China
KW - equilibrium analysis
KW - forecast accuracy
KW - Lotka–Volterra model
KW - Lyapunov function
UR - http://www.scopus.com/inward/record.url?scp=84992496430&partnerID=8YFLogxK
U2 - 10.1080/00036846.2016.1240347
DO - 10.1080/00036846.2016.1240347
M3 - Article
AN - SCOPUS:84992496430
SN - 0003-6846
VL - 49
SP - 2428
EP - 2442
JO - Applied Economics
JF - Applied Economics
IS - 25
ER -