Short-Term Scientific Mission

Towards Deriving Incremental Optimization Problems from Batch Specifications of ROAs

Main theme: Problem Modelling and User Experience
Grantee: Maximilian Kratz, Technische Universitat Darmstadt, Darmstadt, Germany
Host: Steffen Zschaler, King's College London, London, United Kingdom

Start date: 2025-09-08
End date: 2025-09-19
Awarded: 2025-06-05
Report approved: 2025-10-31

Description

The proposed STSM aims to explore how to systematically derive incremental optimization problems from batch-based specifications in the context of Random Optimization Algorithms (ROAs). Using two case studies (teaching assistant allocation and nurse scheduling), the project investigates the development of a (semi-)automated method to adapt solutions when small changes to the initially specified problem (e.g., unavailability of certain resources) occur. By combining model-driven software engineering techniques with the ROAR-NET API, the mission seeks to prototype and formalize an approach for incrementalization. The outcomes are expected to support the derivation of incremental versions in the two exemplary batch-based problem domains, with potential generalization to other domains.

Achievements

The STSM had its focus on the incrementalisation of random optimisation problems. Key outcomes were the foundation of the conceptual incrementalisation approach, as proposed in the application. An arXiv pre-print was created to document the developed ideas. The resulting work is accompanied by the implementation of two variants of the teaching assistant assignment example within the ROAR-NET API and the corresponding problem description. Our research also allowed us to identify and document some limitations of the ROAR-NET API, providing future research opportunities. The STSM was the start of future collaborations, because the work will be further developed into two upcoming publications, one of which aims to extend the pre-print.