Constraint-Adaptive Layout Generation: A Flexible Framework to Integrate Real-World Constraints into Factory Planning

The optimization of production layouts is becoming increasingly important as product diversity increases and corresponding planning becomes
more dynamic. In the context of layout planning, companies must utilize available space optimally to maintain their competitive viability. To
find the optimal solution and address numerous constraints, research is increasingly focusing on automated layout generation and optimization.
However, in real-world applications, a multitude of constraints must be considered, which complicates the automated layout generation process.
The type of parameters to be considered is highly flexible, allowing both hard and soft constraints, as well as dynamic constraints such as space
requirements due to buffers or work positions, to arise from the optimization process. Existing approaches often address these aspects only
partially or simplify them to such an extent that their practical applicability is limited. This paper proposes a constraint-adaptive framework that
integrates diverse types of constraints into the layout optimization process. A categorization of relevant real-world constraints is developed and
... mehr
formalized in a model that supports a flexible and extensible representation of constraints. Different methods for representing boundary conditions
are analyzed. The aim is to show how the constraints are considered in an optimization model, for example, as a graph model or as a penalty
function. Consequently, an adaptive optimization mechanism is implemented, integrating classic metaheuristics with functions for handling
dynamic constraints. To assess the efficacy of the proposed method, a case study is conducted in a brownfield scenario, considering the intricate
spatial and regulatory constraints that characterize such contexts. The optimization results are evaluated in terms of layout quality and additional
performance indicators derived from material flow simulations to validate the practical impact. The proposed approach contributes to a systematic
methodology for incorporating realistic planning parameters into algorithmic layout generation, demonstrating its applicability under real
industrial conditions.