More Virtuous Smoothing

Abstract: In the context of global optimization of mixed-integer nonlinear optimization formulations, we consider smoothing univariate functions

f

that satisfy

f (0) = 0

,

f

is increasing and concave on

[0, + i n f t y)

,

f

is twice differentiable on all of

(0, + i n f t y)

, but

f^{'} (0)

is undefined or intolerably large. The canonical examples are root functions

f (w) := w^{p}

, for

0 < p < 1

. We consider the earlier approach of defining a smoothing function

g

that is identical with

f

on

(d e l t a, + i n f t y)

, for some chosen

d e l t a > 0

, then replacing the part of

f

on

[0, d e l t a]

with the unique homogeneous cubic, matching

f

,

f^{'}

and

f^{″}

at

d e l t a

. The parameter

d e l t a

is used to control (i.e., upper bound) the derivative at 0 (which controls it on all of

[0, + i n f t y)

when

g

is concave). Our main results: (i) we weaken an earlier sufficient condition to give a necessary and sufficient condition for the piecewise function

g

to be increasing and concave; (ii) we give a general sufficient condition for

g^{'} (0)

to be decreasing in the smoothing parameter

d e l t a

; under the same condition, we demonstrate that the worst-case error of

g

as an estimate of

f

is increasing in

d e l t a

; (iii) we give a general sufficient condition for

g

to underestimate

f

; (iv) we give a general sufficient condition for

g

to dominate the simple `shift smoothing'

h (w) := f (w + l a m b d a) - f (l a m b d a)

(

l a m b d a > 0

), when the parameters

d e l t a

and

l a m b d a

are chosen `fairly' --- i.e., so that

g^{'} (0) = h^{'} (0)

. In doing so, we solve two natural open problems of Lee and Skipper (2016), concerning (iii) and (iv) for root functions.

Recommendations

Cites work

Cited in

(6)

Describes a project that uses

Uses Software

This page was built for publication: More Virtuous Smoothing