1: \begin{abstract}

2: This paper derives new algorithms for signomial programming, a generalization of geometric 

3: programming. The algorithms are based on a generic principle for optimization called the MM 

4: algorithm. In this setting, one can apply the geometric-arithmetic mean inequality and a 

5: supporting hyperplane inequality to create a surrogate function with parameters separated.

6: Thus, unconstrained signomial programming reduces to a sequence of one-dimensional minimization 

7: problems.  Simple examples demonstrate that the MM algorithm derived can converge to a boundary point

8: or to one point of a continuum of minimum points.  Conditions under which the minimum point is 

9: unique or occurs in the interior of parameter space are proved for geometric programming. Convergence

10: to an interior point occurs at a linear rate. Finally, the MM framework easily accommodates

11: equality and inequality constraints of signomial type.  For the most important special case,

12: constrained quadratic programming, the MM algorithm involves very simple updates.

13: \keywords{arithmetic-geometric mean inequality \and global convergence \and MM algorithm \and parameter separation \and penalty method}

14: % \PACS{PACS code1 \and PACS code2 \and more}

15: \subclass{90C25 \and 26D07}

16: \end{abstract}

17: