Publications errata

Signomial and Polynomial Optimization via Relative Entropy and Partial Dualization

  • Section 5: in the display labeled “(Ex5)”, the second appearance of the coefficient “15” in the defining expression for “g_1(x)” should in fact be “1.5” (a factor 10 smaller). The current expression is based on an incorrect transcription of the problem statement from Python code.
  • Section 5: in the first row of Table 3, the objective value f(x) of the output of Algorithm 1L should be 11.9643, rather than 11.9600. I’m not entirely sure when or how this error made its way into the paper. Some extra notes: (1) by applying a monotonicity analysis it’s easy to show that the optimal objective is in fact 11.9643, (2) the reported objective value in the paper by Rijckaert and Martens is 11.91, but if you actually evaluate the objective function at the point provided by Rijckaert and Martens, you get 11.96392 and a constraint violation of 5e-6.