Botond Kádár

Hierarchy in distributed shop floor control

Abstract This discussion paper examines the role of hierarchy within the domain of shop floor control in manufacturing plants. Specifically, three distinct structures for shop floor control are considered, namely, hierarchical, heterarchical and holonic control systems. Holonic manufacturing is a novel paradigm for manufacturing that tries to merge the best properties of both hierarchical and heterarchical systems, namely, a high and predictable performance with a high robustness against disturbances. The paper identifies the concept of hierarchy in distributed shop floor control as one of the key contributions of holonic manufacturing. In a holonic architecture, the mapping of functions to control units remains flexible, such that decision power can be distributed over these units as needed. Based on the invariants concept of computer science, a modelling framework is proposed to identify the decision making needs of each unit. It can be used not only to describe the behaviour of the system, but also to impose performance and operational guarantees.

An object-oriented framework for developing distributed manufacturing architectures

Management of complexity, changes and disturbances is one of the key issues of production today. Distributed, agent-based structures represent viable alternatives to hierarchical systems provided with reactive/proactive capabilities. In the paper, approaches to distributed manufacturing architectures are surveyed, and their fundamental features are highlighted, together with the main questions to be answered while designing new structures. Moreover, an object-oriented simulation framework for development and evaluation of multi-agent manufacturing architectures is introduced.

Stability-oriented evaluation of rescheduling strategies, by using simulation

The paper proposes a simulation-based evaluation technique for the testing, validation and benchmarking of rescheduling methods. Dynamic and stochastic scheduling problems with possible scheduling environments are presented. Based on our review of the related literature, frequently applied rescheduling approaches for solving dynamic and stochastic job-shop scheduling problems are analyzed. Schedule evaluation techniques, related measures, and simulation-based schedule evaluation solutions are also introduced and categorized. Certain stability-oriented evaluations of periodic and hybrid rescheduling methods are described for both single- and multi-machine (job-shop) cases. Finally, an industrial application of the proposed method is presented.

Approaches to Managing Changes and Uncertainties in Manufacturing

Abstract Todays complex manufacturing systems operate in a changing environment rife with uncertainty. The performance of manufacturing companies ultimately hinges on their ability to rapidly adapt their production to current internal and external circumstances. Two main kinds of approaches to dealing with the enumerated problems are: to enhance the reactivity of traditionally structured (mostly hierarchical) systems by sophisticated new control techniques, and to construct decentralised, distributed systems. Along both of these - overlapping - lines, the authors developed a genetic algorithm (GA) based dynamic scheduler [4] and a distributed, agent-based shop floor control system to be introduced here which relies on the holonic concept. The approaches are realised by using an object-oriented framework for developing and testing new control architectures and algorithms in the manufacturing domain. Finally, the co-operative use of agent-based distributed control structures and the more centralised (GA-based) scheduler is proposed aiming at systems which can handle critical complexity, reactivity, disturbance and optimality issues at the same time

Adaptation and Learning in Distributed Production Control

Abstract Distributed (agent-based) control architectures offer prospects of reduced complexity, high flexibility and a high robustness against disturbances in manufacturing. However, it has also turned out that distributed control architectures, usually banning all forms of hierarchy, cannot guarantee optimum performance and the system behaviour can be unpredictable. In the paper machine learning approaches such as neurodynamic programming and simulated annealing are described for managing changes and disturbances in manufacturing systems, and to decrease the computational costs of the scheduling process. The results demonstrate the applicability of the proposed solutions, which can contribute to significant improvements in system performance, keeping the known benefits of distributed control.

Towards adaptive and digital manufacturing

The problem each manufacturer repeatedly faces is how to meet demand by making available the required quantities of products in the specified quality and at proper time. From the four main R&D directions, i.e., adaptive manufacturing, digital manufacturing, knowledge-based manufacturing, and networked manufacturing, emphasized by the European initiative Manufuture, mostly the first two are underlined in the paper by illustrating some solution approaches. However, all the related issues, i.e., knowledge-based manufacturing and networked manufacturing, together with the requirements of real-time functioning and cooperativeness are kept in view.

Reinforcement learning in a distributed market-based production control system

The paper presents an adaptive iterative distributed scheduling algorithm that op- erates in a market-based production control system. The manufacturing system is agenti¯ed, thus, every machine and job is associated with its own software agent. Each agent learns how to select presumably good schedules, by this way the size of the search space can be reduced. In order to get adaptive behavior and search space reduction, a triple-level learning mechanism is proposed. The top level of learning incorporates a simulated annealing algorithm, the middle (and the most important) level contains a reinforcement learning system, while the bottom level is done by a numerical function approximator, such as an arti¯cial neural network. The paper suggests a cooperation technique for the agents, as well. It also analyzes the time and space complexity of the solution and presents some experimental results.

Agent-based control of manufacturing systems

Management of complexity, changes and disturbances is one of the key issues in production today. Distributed, agent-based (holonic) structures represent viable alternatives to hierachical systems provided with reactive/proactive capabilities. The paper outlines the difficulties which hinder their industrial acceptance. Several approaches to overcome these barriers are introduced, i.e. the use of simulation techniques for developing and testing agent-based control architectures, the holonification of existing resources and traditional (centralized/hierarchical) manufacturing systems. Finally, the cooperative use of agent-based distributed control structures and the more centralized (e.g. GA-based) schedulers is proposed aiming at systems which can handle critical complexity, reactivity, disturbance and optimality issues at the same time.

Digital enterprise solution for integrated production planning and control

Digital enterprise technologies combined with sophisticated optimization algorithms can significantly contribute to the efficiency of production. The paper introduces a novel approach for integrated production planning and control, with the description of the mathematical models and solution algorithms. The deterministic optimization algorithms are complemented by a discrete-event simulation system to assess solution robustness in case of disturbances. The methods are illustrated by describing two prototype systems and by some experimental results obtained in an industry-initiated project.

Real-Life Scheduling Using Constraint Programming and Simulation

Abstract The paper presents an integrated production planner and job shop scheduler system with flexible modelling capabilities and powerful, scalable solution methods. The system generates close-to-optimal production and capacity plans on the medium term, and detailed production schedules on the short-term. However, the constraint-based, deterministic scheduling model can hardly account for all the uncertainties on the shop floor. Hence, we included such factors into a discrete-event simulation model that is applied to evaluate the robustness of schedules in face of various uncertainties