Geometric programming

A geometric program (GP) is an optimization problem of the form

minimizef0(x)subject tofi(x)≤1,i=1,…,mgi(x)=1,i=1,…,p,

where f0,…,fm are posynomials and g1,…,gp are monomials. In the context of geometric programming (unlike standard mathematics), a monomial is a function from ℝ++n to ℝ defined as

x↦cx1a1x2a2⋯xnan

where c>0  and ai∈ℝ. A posynomial is any sum of monomials.

Geometric programming is closely related to convex optimization: any GP can be made convex by means of a change of variables. GPs have numerous applications, including component sizing in IC design, aircraft design, maximum likelihood estimation for logistic regression in statistics, and parameter tuning of positive linear systems in control theory.

Convex form

Geometric programs are not in general convex optimization problems, but they can be transformed to convex problems by a change of variables and a transformation of the objective and constraint functions. In particular, after performing the change of variables yi=log⁡(xi) and taking the log of the objective and constraint functions, the functions fi, i.e., the posynomials, are transformed into log-sum-exp functions, which are convex, and the functions gi, i.e., the monomials, become affine. Hence, this transformation transforms every GP into an equivalent convex program. In fact, this log-log transformation can be used to convert a larger class of problems, known as log-log convex programming (LLCP), into an equivalent convex form.

Software

Several software packages exist to assist with formulating and solving geometric programs.

See also