Simplifying via Reformulation, Approximation, and Relaxation
摘要
As a means of achieving tractability or reducing the computational burden of an optimization problem, this chapter considers reformulations, approximations, and relaxations. We first consider reformulation procedures and then discuss approximation and relaxation techniques. Regarding approximations and relaxations, we discuss linearization procedures and convexification techniques. We note that the latter are more limited in scope than the former. We note as well that most approximation/relaxation techniques can be improved iteratively. We conclude with some remarks of practical significance.