## Convex envelopes of monomials of odd degree.(English)Zbl 1030.90117

Summary: Convex envelopes of nonconvex functions are widely used to calculate lower bounds to solutions of nonlinear programming problems, particularly within the context of spatial branch-and-bound methods for global optimization. This paper proposes a nonlinear continuous and differentiable convex envelope for monomial terms of odd degree, $$x^{2k+1}$$, where $$k\in \mathbb N$$ and the range of $$x$$ includes zero. We prove that this envelope is the tightest possible. We also derive a linear relaxation from the proposed envelope, and compare both the nonlinear and linear formulations with relaxations obtained using other approaches.

### MSC:

 90C29 Multi-objective and goal programming 90C57 Polyhedral combinatorics, branch-and-bound, branch-and-cut
Full Text: