First results from the 4D4L project published as a technical report

As part of the interdisciplinary project “Data- and target-driven sequential decision-making for time-dynamic logistics systems” (4D4L), IMI, IOR and IFL have jointly published first results as a technical report on arXiv. The paper „Context-Aware Synthesis of Optimization Pipelines for Warehouse Optimization“ was authored by Janik Bischoff, Anne Meyer, Uta Mohring, Fabian Dunke, Maximilian Barlang, Özge Nur Subas, Hadi Kutabi, Stefan Nickel and Kai Furmans.

The paper introduces CASOP, a framework for the context-aware synthesis of optimization pipelines for operational warehouse optimization problems. CASOP links warehouse context, algorithm requirements and a taxonomy of subproblems such as item assignment, order batching, picker routing and scheduling. Based on this structure, CASOP automatically identifies applicable algorithms, composes them into valid optimization pipelines, and evaluates their performance according to operational objectives.

In the experimental validation, CASOP was applied to seven benchmark sets covering four problem classes. In total, 1,063,044 valid pipelines were generated and evaluated on 24,704 instances. The results show that CASOP can represent different warehouse contexts and problem settings, while also identifying cases in which sequential pipeline configurations are particularly effective.

The paper is accompanied by the open-source algorithm repository “ware_ops_algos”. To the best of our knowledge, it provides the first modular open-source implementations of established methods for operational warehouse optimization problems.

The publication can be accessed here: https://arxiv.org/abs/2606.26852.