Implementing Soft Time Windows in cuOpt: Any Ideas or Approaches? #852
Replies: 2 comments
-
Beta Was this translation helpful? Give feedback.
-
|
That is correct, the time windows are treated as hard windows. Though there might be cases the solver may not fit everything in by satisfying all TW constraints, and still report an infeasible solution with some TW's violated with minimal impact on the obj value. You may want to consider this as if the solver treated some of the windows soft. Also, you may be able to address this on the modeling side. For instance, for lower priority tasks you may relax the time window yourself, and for higher importance/priority tasks, leave the time windows unchanged (ie. 0 relaxation). |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
After reviewing the available documentation, I haven’t found any clear guidance on incorporating soft time windows in cuOpt. In real-world routing scenarios, it’s common to allow some flexibility in time windows by applying penalties for early or late arrivals, rather than treating them as strict hard constraints.
I’d like to know if anyone has found a way to model soft time windows in cuOpt, either through existing features, workarounds, or alternative approaches. Even solutions that are not fully native or perfect, but provide a reasonable approximation, would be very helpful.
Thank you in advance for any insights!
Beta Was this translation helpful? Give feedback.
All reactions