动态过程优化算法的计算量比较

A comparison of arithmetic operations of two dynamic process optimization approaches called quasi-sequential approach and reduced Sequential Quadratic Programming (rSQP) simultaneous approach with respect to equality constrained optimization problems is presented. Through the detail comparison of arithmetic operations, it is concluded that the average iteration number within differential algebraic equations (DAEs) integration of quasi-sequential approach could be regarded as a criterion. One formula is given to calculate the threshold value of average iteration number. If the average iteration number is less than the threshold value, quasi-sequential approach takes advantage of rSQP simultaneous approach which is more suitable contrarily. Two optimal control problems are given to demonstrate the usage of threshold value. For optimal control problems whose objective is to stay near desired operating point, the iteration number is usually small. Therefore, quasi-sequential approach seems more suitable for such problems.