You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In this dissertation only minimisation of the maximum completion times for all tasks, commonly referred to as makespan, will be considered, thus ignoring all due-date constraints.
After painstaking parameter tuning, only complex models managed to achieve high training accuracy. Alas, those complex models were severely overfitted to the training instances -- a simple linear model would suffice with similar performance, and for much less overhead! Also, linear models come with the added benefit of easy interpretability.
It's the belief of the author, that the methodology of going about this can be applied to any kind of optimisation problem which involves sequential decision making. As such, then it's suitable to name the framework:
Source
Some rephrasing is needed on these sentences: