Ercan Oztemel

Dynamic job-shop scheduling using reinforcement learning agents

Static and dynamic scheduling methods have attracted a lot of attention in recent years. Among these, dynamic scheduling techniques handle scheduling problems where the scheduler does not possess detailed information about the jobs, which may arrive at the shop at any time. In this paper, an intelligent agent based dynamic scheduling system is proposed. It consists of two independent components: the agent and the simulated environment. The agent selects the most appropriate priority rule according to the shop conditions in real time, while simulated environment performs scheduling activities using the rule selected by the agent. The agent is trained by an improved reinforcement learning algorithm through the learning stage and then it successively makes decisions to schedule the operations.

Intelligent Quality Systems

This study discusses Quality Systems employing techniques from the field of Artificial Intelligence (AI). It focuses upon expert systems and neural networks, two of the most popular AI techniques. Expert Systems encapsulate human expertise for solving complex problems. Neural Networks are able to learn problem solving from examples. The authors illustrate applications of these techniques to the design and operation of effective quality systems. Readers with a background in quality engineering and manufacturing will be able to learn about the uses of expert systems and neural networks to achieve intelligent Statistical Process Control, monitor processes and detect incipient faults in them, design experiments and predict performance, inspect products and monitor and diagnose plants and processes. Readers with an AI background will find a wealth of ideas for practical problems on which to deploy and test their techniques.

An Integrated Neural Network and Expert System Tool for Statistical Process Control

Control charts are a basic means for monitoring the quality characteristics of a manufacturing process to ensure the required quality level. They are used to track product and process variations through graphical representation of the quality variable of interest. A control chart shows the state of control of a process and can exhibit different types of patterns which are indicative of long-term trends in it. This paper describes the integration of an expert system and a neural-network-based pattern recognizer for analysing and interpreting control charts. The expert system has an on-line process monitoring package to detect general out-of-control situations and a diagnosis module to suggest corrective actions. The pattern recognizer is an on-line system comprising two neural networks and an heuristics module designed to identify incipient process abnormalities from control chart patterns. The paper also compares neural networks and expert systems and provides the rationale for the integration process.

A new algorithm for automatic knowledge acquisition in inductive learning

The objective of this study is to present a new algorithm, REX-1, developed for automatic knowledge acquisition in Inductive Learning. It aims at eliminating the pitfalls and disadvantages of the techniques and algorithms currently in use. The proposed algorithm makes use of the direct rule extraction approach, rather than the decision tree. For this purpose, it uses a set of examples to induce general rules. Using some widely used set of examples such as IRIS, Balance and Balloons, Monk, Splice, Promoter, Lenses, Zoo, and Vote, our algorithm is compared with other well-known algorithms such as ID3, C4.5, ILA, and Rules Family.

A general framework for SERM (strategic enterprise resource management)

Enterprise resource planning (ERP) is one of the driving factors for the success of the enterprises. However, recently it was pointed out that companies need resource management rather than resource planning (www.imti21.org). By definition, management involves planning and other capabilities such as monitoring, controlling, organising etc. Due to complexities, dynamic changes and uncertainties in the environment, managers need to define effective strategies allowing their teams to be capable of dealing with unexpected situations. This study presents a new approach called strategic enterprise resource management (SERM) providing facilities for resource management taking strategic planning, customer relations as well as performance management into account.

Control Chart Pattern Recognition

As discussed in the previous chapter, control rules are used to detect out-of-control situations by considering the very recent history of a process. However, to avoid such situations, it is necessary to monitor the long-term history as recorded in control charts. Patterns of variations in a control chart can reveal impending out-of-control situations and help to form cause-effect relationships to predict possible abnormalities in a manufacturing process. In this chapter, automatic control chart pattern recognisers utilising heuristic rules and neural networks as well as combinations of these techniques are described. Experimental results show that these systems are capable of identifying patterns and providing early detections of abnormal conditions with a high degree of accuracy.

Decision support systems in production planning and control

Production planning and control is facing more and more challenging tasks every day. Products are becoming more complex, manual systems are being replaced with complex machines. The world is undergoing a technology revolution and knowledge systems are becoming more dominant then ever before. It seems that the information age will create sophisticated systems requiring complex decisions based on the knowledge of manufacturing and other aspects of organizations. Moreover, traditional manufacturing environments are converging into knowledge-based manufacturing in knowledgeable societies. This obviously indicates the importance of decision support systems which can be developed in such a way that they can utilise knowledge and handle knowledge sources as effectively as possible. The manufacturing industry of the twenty-first century will be characterised by intensive knowledgebased systems of concurrent engineering based on digitalisation, computer network, artificial intelligence, etc. In the coming years, knowledge, agility, intelligence and rapid response are essential requirements for manufacturing systems to favour high quality products, small batch sizes, customer requirements, and environmental consciousness. Taking this into account, the 4th Intelligent Manufacturing Systems Symposium involved a series of papers discussing the decision support which could be created by intelligent systems. Academia and industrial practitioners came together and exchanged their knowledge and experiences with intelligent support systems. The Symposium covered a wide range of manufacturing topics including designing new products, automated storage and retrieval systems, competitive manufacturing strategies and manufacturing knowledge management. The Symposium was a successful event which yielded a series of valuable research publications and discussions. After the discussions, and recommendations by the session chairmen, several papers were nominated to be published in Production Planning & Control. After an extensive review process the papers published in this issue were selected. The papers present results of the studies from improvement through decision making in design to strategic enterprise resource management, from supplier selection to multi agent based simulation and from multi channel scheduling to web-based product development processes. The Symposium will continue biannually to generate a knowledge exchange atmosphere concerning the emergence of new and current technologies creating value for the manufacturing society.

KB-SCHED: Knowledge-based scheduler for complex and dynamic systems

Scheduling problems are becoming more and more complex everyday. This makes the current rules and algorithms difficult to comply with the requirements. New machines with the capabilities of processing more than one jobs is being developed. Sometimes one job is divided into parts and processed by more than one machine at the same time. These make the current algorithms insufficient. Artificial intelligence technologies, especially expert systems are proven to deal with such dynamic complex problems in several domains. In this study, an example of such a complex problem is introduced and knowledge-based scheduling for these kind of problems is elaborated with a real life industrial example.

Technology Readiness Model for Enterprises

Publisher Summary Measuring the technological assessment is one of the main concerns in technology management activities of the enterprises. This chapter presents an innovative technology management model for the enterprises. It provides an overview of existing technology assessment models and introduces the technology readiness model (TRM). TRM is intended to provide a systematic approach to measure the level of technological readiness of the enterprises. It assesses the technology from operational, tactical, and strategic aspects. The model can be implemented in both small and medium enterprises (SMEs) and big companies if business specific modifications can be made. These modifications can be done using artificial intelligence techniques such as expert systems. Neural networks can be used to define the weight factors. The technological readiness levels can be calculated and some remedies can be provided using expert systems as well.

Artificial Intelligence Tools

This chapter reviews expert systems and neural networks, two of the main artificial intelligence tools that have been applied to quality control tasks in manufacturing. The chapter first discusses the basic components of an expert system and briefly surveys expert system development tools including various commercially available shells and environments. The fundamentals of neural networks are then described. Dififerent neural network architectures and learning strategies are detailed. Examples of applications will be presented in subsequent chapters of the book.