A Rollout-Based Algorithm and Reward Function for Efficient Resource Allocation in Business Processes
Jeroen Middelhuis, Zaharah Bukhsh, Ivo Adan, Remco Dijkman · 2025
Read A Rollout-Based Algorithm and Reward Function for Efficient Resource Allocation in Business Processes by Jeroen Middelhuis, Zaharah Bukhsh, Ivo Adan, Remco Dijkman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-11250.txt.
Description
Jeroen Middelhuis and co-authors propose a learning algorithm that assigns workers to business tasks more efficiently, in this 2025 paper. Good assignment shortens the time a case takes from start to finish. Deep reinforcement learning can learn such policies from reward signals, but business settings shift fast. Hand-built reward scores can also drift from the true goal and mislead the learner.
Their rollout method judges each action by simulating the paths that follow it, then improves the policy step by step. The reward comes straight from splitting the cycle-time goal, so no hand tuning is needed. On six test cases with known best policies, the method found the optimum. On larger realistic process models, it matched or beat the best rule-based methods.
Catalog record
| CREATOR | Jeroen Middelhuis, Zaharah Bukhsh, Ivo Adan, Remco Dijkman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-11250