Minimizing rational functions by exact Jacobian SDP relaxation applicable to finite singularities (Q742119): Difference between revisions

The problem of minimizing a rational function subject to equality and inequality polynomial constraints is reformulated, by a standard homogenization technique, as an equivalent (under suitable assumptions) polynomial optimization problem. To solve the latter problem, the authors consider the Jacobian SDP relaxation proposed by \textit{J. Nie} [Math. Program. 137, No. 1--2 (A), 225--255 (2013; Zbl 1266.65094)] and show that this relaxation is exact under an assumption weaker than the one considered by Nie. They also provide some numerical examples.