Note: Execute software for MATLAB software
(2) Based on the interval number DEMATEL method to verify the causal relationships among the four variables in the theoretical framework preliminarily, the differential evolution algorithm is further adopted. Based on the collected sample of 132 teams and 520 team members, this study searches for outstanding individual samples globally, so that the excellent sample which will enter the next generation of population can be used as empirical sample of SEM based on Bayesian estimation. The main purpose of the differential evolution algorithm is to find out the optimal solution vector, which can search for the excellent individual sample with good fault tolerance and strong learning ability as a whole, so that the better individuals with strong learning ability can enter the next generation group, and thus play the optimization and global search capability (Liyanan etc., 2016; Guohaixiang etc., 2014; Deb, 2000). Firstly, based on the differential evolution algorithm, this study uses MATLAB and Stata software to set the execution parameters (as shown in Table 4), and searches for excellent individual sample with good fault tolerance and strong learning ability by the whole, so that the better individuals with strong learning ability can enter the next generation group(Liyanan etc., 2016; Guohaixiang etc., 2014; Deb, 2000). This ensures that the excellent sample which will enter the next generation of population can be used as empirical sample of SEM based on Bayesian estimation exert to optimization and global search function, and explore the causal relationships among the four variables. Secondly, this study reveals the mechanism among the simmelian ties, high performance work practice system, and knowledge fermenting. Finally, the transmission mechanism of the simmelian ties on the knowledge spiral with the high performance work practice system and knowledge fermenting as the conduction variables are revealed.
The relevant principles, modeling steps, and processes of the differential evolution algorithm are as follows(Liyanan etc., 2016; Guohaixiang etc., 2014; Deb, 2000):
The differential evolution algorithm includes four basic operations of initializing population, variation, cross, and selection. The modeling steps and processes of the differential evolution algorithm are as follows(Jinweijian,Huhanhui, 2011; Gaopeiran,Luxinyuan, 2014;Jiangqian etc., 2016):