Software-Framework zur Simulationsbasierten Optimierung mit Anwendung auf Produktions- und Lagerhaltungssysteme (English shop)

Preview

Table of Contents, Datei (94 KB)
Extract, Datei (200 KB)

Michael Kämpf’s dissertation presents the development of a modular software framework for simulation-based optimization of complex production and inventory systems. The research focuses on integrating mathematical optimization methods with computer-based simulation to improve operational decision-making processes. A model-supported optimization approach is introduced, enabling structured representation and flexible adaptation of different simulation models. The proposed hierarchical modeling structure distinguishes between model declaration and model definition, supporting improved reusability and configuration. Particular attention is given to the simulation of stochastic processes and the consideration of random influences in economic and industrial systems. Various heuristic optimization techniques are investigated, including genetic algorithms, tabu search, and simulated annealing. Hybrid, distributed, and parallel optimization methods are additionally examined to reduce computational effort and improve optimization efficiency. Based on these concepts, the object-oriented software framework CAOS is developed and implemented using the C# programming language. Its modular architecture integrates components for model representation, simulation, evaluation, and optimization of different system configurations. Practical applications involve production and inventory models incorporating multiple products, manufacturing units, and capacity constraints. The study evaluates alternative lot-sizing, production release, and sequencing strategies while considering production, inventory holding, and setup costs. Extensive computational investigations are conducted to compare optimization approaches and determine their suitability for minimizing overall system costs. The findings demonstrate that optimization performance depends on the specific model conditions and highlight the potential of adaptable simulation and optimization frameworks for production planning.