MILP Model Comparison

ユーザーの入力を待っています。

Lorenz Wagner

2023年07月23日 21:12

• 
• 
• 

Hello, I have a question. I am currently setting up two shift schedule optimization models. One of them constrains the number of worker shift changes using the 'changes' parameter. In the second model, it is assumed that a worker loses performance per shift change, but theoretically an infinite number of shift changes can be modeled. The performance drop is linear and is controlled by 'alpha'. Now I compare both models using understaffing. In the first model this results inevitably because of the shift-change restriction, in the second model the performance loss is equivalently considered as understaffing. I have now examined how the understaffing behaves in both models when the two parameters are varied. Based on the information I have then calculated a crossover point, but which values which model performs better. Unfortunately I don't know now if this is a suitable idea to compare models. Are there other ways to compare models?

I also did Spearman and Kendall, as well as an ANOVA to test from Significance of the correlations. Do you have any ideas?

関連

• Scheduling

0

• 正式なコメント

  

  Simranjit Kaur

  • Gurobi Staff

  2026年11月13日 01:40

  This post is more than two years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum, or try Gurobot, our chatbot interface offering instant, expert-level support.
• 

  Mario Ruthmair

  • Gurobi Staff

  2023年07月25日 13:59

  Hi Lorenz,

  Are you comparing the performance in terms of solution quality or in terms of solver runtime?
  If you have models that give different optimal solutions since they model the shift changes in different ways, you can hardly compare the two models, only with respect to the solution quality in practice.

  Best regards,
  Mario

  0

投稿コメントは受け付けていません。