Abstract—Advantages and disadvantages in a sequence of unconstrained optimization method are basically compared. The original constrained problem is replaced by a sequence of unconstrained sub-problems through the augmented Lagrangian multiplier method. The unconstrained sub-problems are solved by BFGS method and the sub-direction parallel search quasi-Newton algorithm. Efficiency of this method is compared. The results of numerical tests show that the calculation time of the sub-direction parallel search quasi-Newton algorithm is short and it can solve engineering optimization problems completely.
Index Terms—Quasi-Newton equation, taylor series, global convergence, iteration formula.
The authors are with Ningbo University of Technology, School of Science, China (e-mail: 864787886@ qq.com).
[PDF]
Cite: Wang Bao and Wang Yanxin, "Sub-Direction Parallel Search Quasi-Newton Algorithm," International Journal of Modeling and Optimization vol. 8, no. 3, pp. 131-137, 2018.