解约束优化问题的SQP(Sequential Quadratic Programming)方法探讨
Discussion on SQP (Qequential QuadraticProgramming) Method for SolvingConstraint Optimization Problems
摘要: 本文详细介绍SQP (sequential quadratic programming)方法的设计过程与具体算法,并通过解决等式约束优化问题和不等式约束优化问题的例子来讨论其优劣性,探讨它的扩展性。SOP算法的设计核心是将原来一般约束问题的解转换为一系列简单子问题(例如二次规划问题)的求解。而二次规划问题的求解方法非常成熟与完善,这种将复杂优化问题的求解转化一系列简单问题的求解方法被提出之后广受欢迎,使得SQP方法成为解决非线性约束优化最有效的方法之一。它在解决具有非线性优化问题时有它独特的优势。理解和掌握它,对于理解与应用其它的SOP方法也有极大的帮助。
Abstract:
This paper introduces the design process and specific algorithm of SQP (sequential quadratic programming) in detail, and discusses its advantages and disadvantages and its expansibility by solving an example of equality constraint optimization problem and inequality constraint optimization problem. The core of SQP algorithm is to convert the original solution of general constraint problem into a series of simple subproblems (such as quadratic programming problem). However, the solution method of quadratic programming problem is very mature and perfect. This method, which transforms the solution of complex optimization problems into a series of simple problems, has been widely popular after being proposed, making SQP method one of the most effective methods to solve nonlinear constrained optimization. It has its unique advantages in solving nonlinear optimization problems. Understanding and mastering it is also of great help to understand and apply other SQP methods.
参考文献
[1]
|
孙中波,段复建.不等式约束优化的非单调可行信赖域-SQP算法[J].应用数学学报,2011,34(4):655-670.
|
[2]
|
孙中波,段复建,许春玲,田彦涛.不等式约束优化超线性收敛的信赖域-SQP算法[J].应用数学学报,2014,37(5): 878-890.
|
[3]
|
王华.解等式约束规划的信赖域SQP滤子方法[J].内蒙古师范大学学报(自然科学汉文版),2008,37(1): 1-5.
|
[4]
|
王珺.非线性等式约束优化问题的信赖域滤子算法研究[D]:[硕士学位论文].太原: 太原科技大学,2010.
|
[5]
|
周敏.约束非线性优化的信赖域滤子SQP算法[D]: [硕士学位论文].洛阳: 河南科技大学,2014.
|
[6]
|
李董辉,童小娇,万中.数值最优化算法与理论[M].北京: 科学出版社,2010.
|
[7]
|
Nocedal, J. and Wright, S.J. (1999) Numerical Optimization. Springer, Berlin. https://doi.org/10.1007/b98874
|