1
Paper presented at Industrical Engineering Conference, Three Gorges, China, 1999
The Human Side of Holonic Manufacturing Systems
Hongyi SunPatri K V and Patri K. Venuvinod and Hongyi Sun
Department of Manufacturing Engineering & Engineering Management
City University of Hong Kong, Kowloon, Hong Kong
Abstract Holonic manufacturing systems were proposed to cope with the dynamic environment in manufacturing. However, majority of research on this topic is on technical side. This paper will review the concept of holonic manufacturing systems and discuss the human side of the system. An example of a holonic work organization will be described. A management framework for design and implement holonic work organizations will be proposed. Implications for practices and future research will be explored.
Key words: Holonic manufacturing, work organization, future manufacturing
1 Introduction
It may be recalled that, during the 1980s, the concepts of Computer Integrated Manufacturing (CIM) and Flexible Manufacturing Systems (FMS) had generated much euphoria in the 1980s. Such enterprises were touted as the ‘Factories of the Future’. Egged on by this euphoria many countries (including China) embarked on the path of promoting CIM and establishing FMSs. How wise have they been? It appears that, while the faith in CIM was well placed, that in FMS was premature. The ubiquity of computers has validated the goal of CIM. The recent growth in Internet and globally distributed manufacture have now turned CIM almost into an imperative. However, one cannot say the same about FMS with any confidence. FMSs, as implemented so far, are expensive since they require that all hardware components be computer controlled and be able to freely communicate each other. Thus, an FMS is, by nature expensive. Yet most FMSs failed to deliver the promised results. Why? The first reason is that most of the FMSs were developed and run in a hierarchical manner without realizing the system dynamics was too complex to be anticipated and accounted for during the design phase. The second reason was the misplaced view that, since an FMS is an automated entity, it was sufficient to focus on the technological aspects of the systems.
Following the abatement of the euphoria about FMSs and, in response to rapid changes in the dynamics of global markets and the trend towards in increasing customer requirements for higher product variety, To cope with the dynamics in markets and changes in customers' requirement for varieties of new products, companies around the world had have been started looking for new paradigms of manufacturingthat provide increased which have both flexibility as well as nd productivity. In this process, Computer Integrated Manufacturing (CIM ) has continued to be been promoted as the strategy concept that is expected to will somehow deal with the challenge. However, as Warneckes (1993) points out, CIM often wrongly assumes that the organizationalstructure and procedures can be automatically improved by the implementation of CIM. Since this is not true, often,As a result,problemsarisehaveduring been identified through the process of CIM implementation (Suda, 1989, Youssef 1990, and Sun 1996, etc.). AlEven tthough more and more companies are trying to convert into what has been termed as "agile manufacturing enterprises" ( Montgomery and Levine, 1996), only very little work has been done towards establishing a theoretical framework for autonomous and decentralized manufacturing organizations (Tonshoff, Winkler and Aurich, 1994). With a view to filling this critical gap,In 1993 and 1994, a world wide feasibility study called Intelligent Manufacturing Systems (IMS) involving for an international collaborative research program in manufacturing was initiatedundertaken in the period 1993-1994, called Intelligent Manufacturing System (IMS). IMS aims at a joint systemization and standardization ofor the manufacturing technologies of the 21st centuryies. Holonic Manufacturing System (HMS) is one major paradigm being developed and tested as a part of of the test case in IMS (Van Brussel et al, 1998).
HMS is inspired by the concept of “holon”concept introduced by Koestler (198967) in his work on biological nic and social systems. The termword holon is a result of the fusion of the Greek wordholos meaning whole and a suffix -on suggesting that it is a part of a larger systems. Thus, the concept of hHolon conceptincludes the consideration of describes the relationship between the whole and its the parts. At first glance, this seems similar to the relationship between a system and its sub-systems in general system theory (Skyttner, 1996).
The holonic concepts described above At first glance, this may seem It is similar to the relationship between a system and its sub-systems in general system theory (Skyttner, 1996). However, tThe concept inspired IMS researchers in IMS to propose athe conceptual framework for intelligent manufacturing, identified as holonic manufacturing systems (HMS) (Tonshoff, Winkler and Aurich, 1994). The distinguishing characteristic of a HMS is that holons must exhibit the characteristics ofcomposed of autonomy and cooperation. Note that, being autonomous, holons are similar to ‘agents’ which are being increasingly adopted by software developers. ous and cooperative holons or agents. The details of HMS will be reviewed in the next section.
It may be recalled that during the 1980s much euphoria was generated
Although the intention of HMS is to establish a theoretical framework for the new organizationalstructure of future manufacturing system, much majority of the research on HMS has focused on is on the technological ical side whereasan thed human and organizational sides have are eitherbeen ignored or emphasizedto a much smaller extentlittle. Previous research proved that, at present, if any the best manufacturing model were to succeed in the real world it must is still the ccombineationthe of technologiical and human elements and, in this respect, the concept of CIM is far from being developed,while fully computer integrated manufacturing system is still a future issues (Yousseff, 1992, Sun 1996). As the experience worldwide (including that derived from mainland projects) shows, most CIM/FMS projects fail not because of a lack of understanding of the technologies involved but due to the inability to effect the required changes in organizational structures and culture. Thus, it is now generally accepted that human factors must be taken into consideration while designing To design and implementing a balanced manufacturing system, human factors must be taken into consideration. The goal of the present paper is to highlight certain issues that need consideration in such an endeavor. In particular, it is intended The intention of this paper is to 1) introduce the basic concept of holonic manufacturing system, 2) describe an example holonic work organization, and 3) discuss propose a suitable management framework for holonic work organizations.
2 Review of Holonic manufacturing system
It can be argued that The holonic approach recognizes that the manufacturing enterprise involvesequipment with varying degrees of automation. It does not depend on a hierarchical command structure and recognizes the principleof self-organization. It provides an enterprise- view based on business processes aligned to a client-server architectural framework and a robust approach for dealing with changes to organization and processes (Thompson and Hughees, 1998). The HMS consortium developed the following list of definitions to help understand and guide the translation of holonic concepts into a manufacturing setting (Van Brussel et al, 1998) as shown in Ffigure 1.
Figure 1 An illustration of holons (adaptedSimplified from,Tonshoff, Winkler and Aurich, 1994)
2.1 Holon
Figure 2 shows the general structure of a holon (Yuen, 1999).
Figure 2 The general structure of a holon (Yuen, 1999)
A holon by itself is a functional unit, i.e. it is capable of performing a certain function that serves the overall system purpose. This ability to perform its function is represented by the ‘Function Shell’ in Figure 2. In the context of certain functions, the corresponding holon may include a physical processing part. However, since the concept of HMS has been conceived in the context of a computer integrated environment, every holon must consist of an information processing part. Thus ‘Data Storage’ (see Figure 2) is at the very core of a holon.
The ‘Function Shell’ and ‘Data Storage’ components are not unique to the concept of holonic manufacturing systemsthey are recognized in any type of computer integrated system including the traditional hierarchically organized FMSs. However, what makes the holonic paradigm unique is the expectation that every holon exhibits the characteristics of autonomy and cooperation. Thus a holon may be defined as aAn autonomous and co-operative building block of a manufacturing system for transforming, transporting, storing and/or validating information and physical objects. AThe holon consists of an information processing part and often a physical processing part. A holon can contain smaller holons and can be part of a largerbigger holon as shown in Ffigure 1. Each production unit (e.g., a machine) can be a holon and the variousseholonsunits cooperate with each other-from planning and scheduling to physical production-to manufacture products.
Autonomy and cooperation capabilities of individual elements are the most important features to be considered while for developing such distributed and decentralized decision making architectures. The triangular elements at the four corners of Figure 1 represent the additional abilities that a holon needs to possessed if it were to be able to act autonomously and cooperatively.
2.2 Autonomy
Autonomy is tThe capability of an entity to create and control the execution of its own plans and/or strategies.
A holon self-regulates (although it does not self-organize) and reacts to changes in the environment and the changes are fed back to the centresystemcontrolling it. Hence, a holon needs to be ‘self aware’, i.e. be aware of its own capabilities so that it can participate in serving the overall system purpose in an opportunistic manner. Thus, a holon is expected to perform ‘Internal Coordination’see Figure 2.
It may be reiterated that tThe autonomy of holons is are not absolute and is subject to the goals and objectives of the overall system. Thus, a holon is only at is why it is should relatively autonomous ( Mathews, 1995).
2.3 Co-operation
Holons co-operate using peer-to-peer behavior for the exchange of both information and physical components. Each holon continuously determines its role and position through interaction with the other holons. Hence a holon is expected to have the able to co-ordinate through dynamic communications with other holons (superiors, peers, or subordinates) as appropriatesee Figure 2. Thus, unlike units in a hierarchical system where the command hierarchies are fixed, hHolons can enter temporary command hierarchies if the system goals that the which the holon knows can thereby be achieved easier and more efficiently.
The ability of a holon to make its own decision constitutes its autonomy whereas the property of trying to comply in the first instance with the suggestions (which would have been treated as orders in a purely hierarchical system) received ensures the co-operation feature 9Yuen, 1999).
2.4 Holarchy
A hHolarchy is a system of holons that can co-operatively interacte to achieve the a goalsand or objectives of the overall system. The holarchy defines the basic rules for co-operation of the holons and thereby limits their autonomy. The holarchy is the architecture in which holons are organized and co-operate with each other. In the holarchy, holons are dynamically combined with each other in order to adapt themselves to given objectives and environments. It is inherently a lean organization without static hierarchies.
However, an A HMS is not organized in a fixed way,abut organizes itself dynamically to meet its goals and adapts itself to changes in its environment or itself. Thus itIt can also organized itself in terms of temporary hierarchies. Within such a temporary hierarchy the holons are self-regulating parts which function (Tonshoff, Winkler and Aurich, 1994):
- as autonomous wholes in super-ordination to their parts,
- as dependent parts in subordination to higher level goals,
- in coordination with their local environments (Tonshoff, Winkler and Aurich, 1994)..
Some principles that guide the design of a holarchy are available in (Yuen, 1999).
2.5 Holonic SystemsSummary
AIn summary, HMS is a holarchy of holons. A hoThe holons can be a machineing tool, a robot, a planning unit or an information processing centere. Bruseel et al (1998) proposed three basic holons: order holon, product holon and resource holon. They agree that staff holon can be added. However, staff holon only play the role of assisting the three basic holons with expert knowledge. Staff holons are not described in details. According to Tonshoff, Winkler and Aurich (1994), whilst a human participant is are often a part of a holon, they are not always required. If (s)hetheyisare a part of a holon, (s)hetheyisare most likely considered part of the information processing and sometimes also the physical processing part. Deen (1993) claimed that an HMS is a new concept in manufacturing, in which "intelligent" autonomous units called holons co-operate together to provide a flexible but largely unmanned environment for manufacturing. These studies indicate that the current research on HMS, as in previous research on CIM, is based on the assumption that the future factory will be unmanned and, as a result, will be overwhelmingly dominated by on the technical details of intelligent applications of computer aided technologies, especially on the detailed scheduling level (Gou, Peter and Kyoya, 1998; Sousa, and Ramos, 1999). Little work seems to have been done towards has been done on probing the organizational foundations (Mathews, 1995). In the next section, a holonic work organization will be described with a view toof attracting the attention of researchers on human and organizational sides to of HMS.
The goal of HMS is "to attain in manufacturing the benefits that holonic organization provides to living organisms and societies, i.e. stability in the face of disturbances, adaptability and flexibility in the face of change, and efficient use of available resources." However, living organisms and societies are by nature extremely complex. Koestler's tried to explain how such extremely complex systems are able to function. He suggested that such complex systems evolve from simple systems much more rapidly if there are stable intermediate forms than if they are not (Valckenaers, 1997). For instance, life emerged billions of years ago from some warm little pond through random interactions among its simple molecular building blocks. Somehow these interactions produces an autocatalytic set: a set of elements that catalyzed the creation of its own elements which, in turn, rapidly populated the pond. Autonomy and cooperation are the hallmark features of these autocatalytic elements although the presence of these features by itself does not guarantee autocatalycism. For instance, a modern manufacturing system comprises of three types of entities: hardware, software, and humanware. Each of these types can be made to possess the properties of autonomy and cooperation. However, whereas it is easy to accept that humanware can be autocatalytic and there are signs of software exhibiting autocatalycism, it is difficult to see how hardware could be autocatalytic (although, in some distant future, robots might design and build other robots).
The holonic paradigm has already been tested in several manufacturing contexts by IMS teams. For instance, an flexible assemblysystem has been operated in hierarchical, heterarchical and holonic modes at Katholieke Universiteit Leuven in Belgium and it was found that the holonic mode yielded far superior results in terms of flexibility, adaptability to change, and robustness (Valckenaers 1997). Likewise, when an NC controller was designed following holonic concepts, it was found that the NC holon was more flexible and responsive to interrupts. More interestingly, the NC holon exhibited self-assertive and self-integrative behavior (Kruth et al, 1996)
3 A holonic work organization around a FMS
The case company is a manufacturer of engines for ships (Sun and Gertsen, 1995). It sells itstheir products world-wide and is pretty competitive. In 1986, the milling shop started to use a small Flexible Manufacturing System (FMS). It has two milling machining centeress, each of which has a tool magazine with about 60 tools. There is a tool warehouse that has 200 types of tools supporting behind tthe system which has 200 tools. Initially, a A robot waswas used to exchange the tools both between the magazines and the tool warehouse. There are two loading/unloading stations. The transportation system consisted of is rail and pallets. The pallets wereare sent into the machining centers by another robot. A computer system controlleds all the operations.
At the beginning the system did not work as expected, the utilization rate was lower than 70%, breakdowns wereas too frequentoften and the productivity was not high. It looked as if the problem was technical and much effort was devoted to repair and maintaiin the physical equipment and computer system. Few changes were introduced in the organizational aspect. The performance did not improve.
The project manager himself left the company and moved to another place in 1989. The management found a new technology manager which was also in charge of the FMS system. Since the new manager took over the management of the FMS, a series of organizational changes were implemented, without further investments in technology. And as a result, the system has turned out to be very efficient and successful. The utilization rate of the FMS was increased to more than 90%. The system can now produce more than 60 different milling parts. In the following part, the work organization around the FMS will be described from the holonic point of view.
3.1 The work organization
The new work organization around this system can be characterized as an autonomous and cooperative working group. In this group, there are seven persons, including an implicit foreman. The other six are all skilled workers. Although they have joined training courses given by the vendor, t. The company has decided that every other year, they will be sent to the vendor for a new course. All of them can be classified as skilled workers. Normally, it needs three people to operate the system. The group can be recognized as a holon and the three workers are second-level holons.