Timeline for Linearizing the square root of binary summations
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 20, 2020 at 7:26 | history | edited | TheSimpliFire♦ | CC BY-SA 4.0 |
deleted 24 characters in body
|
| Feb 20, 2020 at 1:36 | history | became hot network question | |||
| Feb 19, 2020 at 18:54 | vote | accept | tcokyasar | ||
| Feb 19, 2020 at 18:27 | answer | added | RobPratt | timeline score: 6 | |
| Feb 19, 2020 at 18:23 | comment | added | tcokyasar | Yes, it is nonnegative, but, sorry I couldn't get what you mean. (I understand the relaxation but cannot figure out how to form the constraints exactly.) | |
| Feb 19, 2020 at 18:22 | comment | added | RobPratt | If the objective coefficient of $z_j$ is nonnegative, you can relax your equality to $z_j \ge$ and then apply one of the transformations in the first link. | |
| Feb 19, 2020 at 18:20 | comment | added | tcokyasar | I agree! $z_j$ appears in the objective function (minimize) with a constant associated coefficient and $x_{ijk}$ and $y_{jik}$ appears everywhere as they are some routing variables for a given $k$. Can we extend the domain of $n$ based on the possible highest summation? | |
| Feb 19, 2020 at 18:17 | comment | added | RobPratt | You cannot interchange the $\sqrt{}$ and $\sum$ like that. Where do $x$, $y$, and $z$ appear elsewhere in the model? | |
| Feb 19, 2020 at 17:38 | comment | added | tcokyasar | I feel like my attempt is not correct because the binary summation inside the square root (excluding $\lambda_k$) is not in the domain of $\{0,1,2\}$. Am I right? If yes, any alternative solution suggestions? | |
| Feb 19, 2020 at 17:28 | history | asked | tcokyasar | CC BY-SA 4.0 |