Torsten Hildebrandt

Towards improved dispatching rules for complex shop floor scenarios: a genetic programming approach

Developing dispatching rules for manufacturing systems is a process, which is time- and cost-consuming. Since there is no good general rule for different scenarios and objectives automatic rule search mechanism are investigated. In this paper an approach using Genetic Programming (GP) is presented. The priority rules generated by GP are evaluated on dynamic job shop scenarios from literature and compared with manually developed rules yielding very promising results also interesting for Simulation Optimization in general.

On using surrogates with genetic programming

One way to accelerate evolutionary algorithms with expensive fitness evaluations is to combine them with surrogate models. Surrogate models are efficiently computable approximations of the fitness function, derived by means of statistical or machine learning techniques from samples of fully evaluated solutions. But these models usually require a numerical representation, and therefore cannot be used with the tree representation of genetic programming (GP). In this paper, we present a new way to use surrogate models with GP. Rather than using the genotype directly as input to the surrogate model, we propose using a phenotypic characterization. This phenotypic characterization can be computed efficiently and allows us to define approximate measures of equivalence and similarity. Using a stochastic, dynamic job shop scenario as an example of simulation-based GP with an expensive fitness evaluation, we show how these ideas can be used to construct surrogate models and improve the convergence speed and solution quality of GP.

Engineering autonomously controlled logistic systems

Today enterprises are exposed to an increasingly dynamic environment. Last but not least increasing competition caused by globalization more and more requires gaining competitive advantages by improved process control, within and beyond an enterprise. Autonomous control of logistic processes is proposed as a means to better face dynamics and complexity. Autonomous control means the ability of logistic objects to process information, to render and to execute decisions on their own. To engineer logistic systems based on autonomous control, dedicated methodologies are needed. This paper proposes a methodology for system specification that consists of a notational part, a procedure model and a software tool, covering a substantial part of the overall system engineering process. Supported by this methodology a logistics process expert will be able to specify an autonomous logistic system adequately. Further research will later on complement the methodology to support the whole engineering process.

Generating dispatching rules for semiconductor manufacturing to minimize weighted tardiness

Dispatching rules play an important role especially in semiconductor manufacturing scheduling, because these fabrication facilities are characterized by high complexity and dynamics. The process of developing and adapting dispatching rules is currently a tedious, largely manual task. Coupling Genetic Programming (GP), a global optimization meta-heuristic from the family of Evolutionary Algorithms, with a stochastic discrete event simulation of a complex manufacturing system we are able to automatically generate dispatching rules for a scenario from semiconductor manufacturing. Evolved dispatching rules clearly outperform manually developed rules from literature.

Hyper-heuristic evolution of dispatching rules: A comparison of rule representations

Dispatching rules are frequently used for real-time, online scheduling in complex manufacturing systems. Design of such rules is usually done by experts in a time consuming trial-and-error process. Recently, evolutionary algorithms have been proposed to automate the design process. There are several possibilities to represent rules for this hyper-heuristic search. Because the representation determines the search neighborhood and the complexity of the rules that can be evolved, a suitable choice of representation is key for a successful evolutionary algorithm. In this paper we empirically compare three different representations, both numeric and symbolic, for automated rule design: A linear combination of attributes, a representation based on artificial neural networks, and a tree representation. Using appropriate evolutionary algorithms (CMA-ES for the neural network and linear representations, genetic programming for the tree representation), we empirically investigate the suitability of each representation in a dynamic stochastic job shop scenario. We also examine the robustness of the evolved dispatching rules against variations in the underlying job shop scenario, and visualize what the rules do, in order to get an intuitive understanding of their inner workings. Results indicate that the tree representation using an improved version of genetic programming gives the best results if many candidate rules can be evaluated, closely followed by the neural network representation that already leads to good results for small to moderate computational budgets. The linear representation is found to be competitive only for extremely small computational budgets.

Uml as a Basis to Model Autonomous Production Systems

This article will investigate the suitability of the Unified Modelling Language (UML) for requirements analysis of autonomous logistic processes by a logistics domain expert. Such a model is the basis for subsequent implementation of the system consisting of software engineering and hardware configuration. Relevant parts of UML will be used to model an exemplary scenario which will form the basis to derive benefits and drawbacks of using the UML in this context. Suggestions on how the identified gaps can be filled will be presented in the paper.

Dynamic adjustment of dispatching rule parameters in flow shops with sequence-dependent set-up times

Decentralised scheduling with dispatching rules is applied in many fields of production and logistics, especially in highly complex manufacturing systems. Since dispatching rules are restricted to their local information horizon, there is no rule that outperforms other rules across various objectives, scenarios and system conditions. In this paper, we present an approach to dynamically adjust the parameters of a dispatching rule depending on the current system conditions. The influence of different parameter settings of the chosen rule on the system performance is estimated by a machine learning method, whose learning data is generated by preliminary simulation runs. Using a dynamic flow shop scenario with sequence-dependent set-up times, we demonstrate that our approach is capable of significantly reducing the mean tardiness of jobs.

Large-scale simulation-based optimization of semiconductor dispatching rules

Developing dispatching rules for complex production systems such as semiconductor manufacturing is an involved task usually performed manually. In a tedious trial-and-error process, a human expert attempts to improve existing rules, which are evaluated using discrete-event simulation. A significant improvement in this task can be achieved by coupling a discrete-event simulator with heuristic optimization algorithms. In this paper we show that this approach is feasible for large manufacturing scenarios as well, and it is also useful to quantify the value of information for the scheduling process. Using the objective of minimizing the mean cycle time of lots, we show that rules created automatically using Genetic Programming (GP) can clearly outperform standard rules. We compare their performance to manually developed rules from the literature.

Specifying Adaptive Business Processes within the Production Logistics Domain — A new Modelling Concept and its Challenges

Today enterprises are exposed to an increasingly dynamic environment. Last but not least increasing competition caused by globalisation more and more requires gaining competitive advantages by improved process control, within and beyond the borders of producing enterprises. One possibility to face increasing dynamics is autonomous control of logistic processes. This shall allow more robust processes in spite of growing environmental as well as internal complexity.

Modeling of orders in autonomously controlled logistic systems

The paper extends the Autonomous Logistics Engineering Methodology (ALEM) by a deeper understanding of immaterial logistic objects to trigger manufacturing processes. Further, a hierarchical modeling concept is introduced to split customer orders logically into partial orders, which run directly at the shop floor level. Each partial order consists of certain manufacturing steps. The amendments enable adequate modeling of autonomous manufacturing processes. The research is a further step to integrate autonomously controlled processes in logistics.