1
$\begingroup$

I have the following question. I have just read this question and was wondering if it is possible to extend this problem. I am interested in whether it is also possible to identify which machine is being switched to which machine? For example, that in $t=2$ on machine $m=3$ the product $p$ is changed to machine $m=2$ in $t=4$, whereby on day $t=3$ the product was not on any machine. For example, by a binary variable $w_{mm't}$, where $m\neq m'$.

$\endgroup$

2 Answers 2

1
$\begingroup$

One way to do this is to define binary variables $x_{p,m,t},$ where $x_{p,m,t}=1$ if and only if as of time $t$ the most recent machine to process job $p$ (if any) was machine $m.$ So in your example you would have $x_{p,3,2}=1$ (job $p$ is processed on machine 3 on day 2), $x_{p,3,3}=1$ (the job is not being processed on day 3, so the most recent machine to process it is still machine 3), and $x_{p,2,4}=1$ (the job is being processed on machine 2 on day 4, so machine 2 is the now the most recent machine). You can then define $w$ via $$w_{p,m,m',t}\le x_{p,m,t-1}$$ $$w_{p,m,m',t}\le x_{p,m',t}$$ $$w_{p,m,m',t}\ge x_{p,m,t-1} + x_{p,m',t} - 1.$$

$\endgroup$
4
  • $\begingroup$ Thanks for your answer. I guess you mean $x_{p,3,2}=0$ correct? $\endgroup$ Commented Jul 15, 2024 at 20:34
  • $\begingroup$ The wording of your example is a bit confusing. I took it to mean that $p$ is on machine 3 on day 2, not on any machine on day 3, and on machine 2 on day 4. If so, then $x_{p,3,2}$ should be 1, signalling that on day 2 $p$ has most recently been on machine 3 (where it currently is). If you meant something else, perhaps you could clarify your example. $\endgroup$ Commented Jul 15, 2024 at 21:27
  • $\begingroup$ I see thatnk you. And how would i define the new $x_{p,m,t}$ based on the post i mentioned in the OP? $\endgroup$ Commented Jul 16, 2024 at 9:31
  • $\begingroup$ See the answer by @themaneater22. $\endgroup$ Commented Jul 16, 2024 at 16:06
1
$\begingroup$

Based on the anwser of @prubin and the answer to the question mentioned in the OP, the anwser using the same variable mentioned there looks like:

$$w_{p,t, m,m'} \leq d_{p,t, m}\quad \forall p\in P, t\in T, m\in M, m'\neq m$$ $$w_{p,t, m,m'} \leq s_{p,(t-1), m'}\quad \forall p\in P, t\in \{2,...,T\}, m\in M, m'\neq m$$ $$w_{p,t, m,m'} \geq (d_{p,t, m}+ s_{p,(t-1), m'})-1\quad \forall p\in P, t\in \{2,...,T\}, m\in M, m'\neq m$$

This ensures, that if there is a change from $m'$ to $m$, this will be identified. Then you can sum up all values of $w_{p,t, m,m'}$ and then the sum will be exactly equal to 1 if there was a change and 0 if there was none.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.