MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

APPLICATION OF PRODUCTION SYSTEM CLASSIFICATIONS IN RAPID DESIGN AND VIRTUAL PROTOTYPING

Christen Rose-Anderssen1, James Scott Baldwin1, Keith Ridgway1, Fabian Boettinger2, Kwabena Agyapong-Kodua3, Ivan Brencsics4, Istvan Nemeth5

1Advanced Manufacturing Institute, The University of Sheffield Sheffield, UK

(MEP09CR, j.baldwin, k.ridgway)@sheffield.ac.uk

2Fraunhofer Institute for Manufacturing Engineering and Automation IPA Stuttgart, Germany

Fabian.Boettinger@ipa.fraunhofer.de

3Precision Manufacturing Centre, Faculty of Engineering, University of Nottingham Nottingham, UK

Kwabena.Agyapong-Kodua@nottingham.ac.uk

4Gamax Kft Budapest, Hungary

ivan.brencsics@gamax.hu

5Department of Manufacturing Science and Technology, Budapest University of Technology and Economics Budapest, Hungary

inemeth@manuf.bme.hu

Abstract:

This paper presents a practical application of classifications of production systems which underlie a web-based expert system for automating the identification, diagnosis and improvement of such systems. This will complement the larger COPERNICO software system, which simplifies and makes accessible essential modelling tools for the rapid design, simulation and virtual prototyping of factories. The objectives are to systematically organise manufacturing systems and their characteristics in classifications to reveal the change process that connects the development of processes and technologies, to their overarching system; and to develop a useful benchmarking tool to enable users to view manufacturing systems in an evolutionary landscape, gauge performance, and identify strategies and tools for change and improvement.

Keywords:

Evolution, Cladistic Classification, Linnaean Classification, Manufacturing Cladistics

1 INTRODUCTION

The aim of this work is to develop a classification system applicable to the wide-ranging and diverse organisations that make up the discrete manufacturing sector of the EU where a wide variety of manufacturing systems and technologies are employed to make a wide range of

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

products. The main objectives are: a) to identify a range of characteristics (characters and states) with which to describe discrete manufactuirng systems; and through b) the development of both hierarchical and cladistic classifications, c) define the most evolutionary relevant characters and states. Therefore, inquiries underlying evolutionary theories of classification can firstly be applied for clarifying existing manufacturing systems. In that sense, taxonomy is the classification in an ordered system that indicates natural relationships based on the suspected evolution of the system under consideration. It can also be applied to identify and discuss emergent systems. Finally, it can be used as a holistic benchmarking and diagnostics tool for discussing and monitoring manufacturing properties in light of strategies for the future. This will complement a larger software system architecture and tools for the rapid design and virtual prototyping of factories of the future (www.copernico.co).

1.1 Science of classifications

Classification, as a science, began with biologists. Taxonomy and classification have been useful tools in managing the information on living entities, their genetics, form and behaviour. The system of hierarchical biological classification was originally described by Carl Linnaeus (later; von Linne) in his book, Systema Naturea originally written in 1735 [23]. Here von Linne describes systematics as the scientific inquiry into biological differences. The group into which organisms are placed are referred to as taxa (singular; taxon). The taxa are arranged in a hierarchy originally limited to Kingdom, Class, Order, Genus, and Variety. The taxon is ranked within this hierarchy, and is the primary concern of relationship investigation. The basic rank of the taxon is the Species. The second-most important rank is that of the Genus followed by the Family and so on. The Linnaean hierarchy, as it ranks entities artificially, can be misleading as it suggests that different groupings within the same rank are equivalent.

Phylogenetic classification, on the other hand, firstly, tells you something important about the organism – its evolutionary history and secondly, does not attempt to rank organisms. Phylogenetic classification (now known as cladistics), constructs classifications by grouping organisms that share derived characters [16]. Camin and Sokal [9] suggest the term cladogram to distinguish a cladistic (Klados being a Greek term for ‘branch’) dendrogram from a phenetic one (phenogram). Cladistics is an evolutionary classification scheme that not only describes the attributes of existing entities but also the ancestral characteristics. A classification starts off by defining the nature of the problem to be solved [22]. In terms of cladistics, this means trying to understand the evolution of the identified classification problem. This assists in setting clear boundaries for which characters are relevant for the selected clade. The clade is a group which consists of the group of Species under study. Each Species is defined by a list of characteristics, called character states in classification theory [31]. These character states distinguish one Species from another. The considered (selected) Species are defined by their possession of derived character states. The resulting cladogram is constructed by grouping Species that share a similarity of change.

A phylogenetic group is made clear though the distinction between homologies and analogies/homoplasies. Homology is present when a character is shared between Species and their common ancestors. Analogy, however, occurs when a character is shared between Species and that character was not present in a common ancestor. Lorenz [25] argues that, independently of each other, two different forms of life may functionally take similar, parallel evolutionary paths in adapting themselves to the same external circumstances. This convergent evolution of a character state between unrelated animals therefore produces a superficial relationship between them and is refered as an analogy or homoplasy [24].

Evolutionary change is about change of one or more characters from one state to another. A realistic situation occurs when characters are in conflict about the relationship among taxa. Then it is important to choose a tree that requires the fewest assumptions about character

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

state changes as it is more parsimonious [24]. The principle of parsimony suggests that all intermediate stages between the recent ancestor and the considered Species possess the same state [9] [37] [39] [45]. The total number of state changes necessary to support the relationship for the Species in a cladogram describes the tree length [32]. The cladogram with the minimum length is therefore considered to be the best-fit tree. The principle of parsimony is the most commonly used principle to estimate phylogeny [22].

The law of Phylogenetic Irreversibility, or Dollo’s Law [15], states that evolution rarely reverts to an earlier specialised stage. This means that when a line has attained a derived character state it cannot return to a state that is ancestral [9] [39]. However, this is a truth that can be subjected to certain modifications. Darwin [11] observes that reversals of characters to ancestral states do occur. McCarthy et al [31] argue that phylogenetic reversal also occurs as part of the selective and adaptive circumstances of evolution in manufacturing. Central to a cladistic analysis is to determine which character states are plesimorphic (primitive) and which are apomorphic (derived). In out-group comparison, if a taxon, which is not a member of the group of organisms being classified, has a character state that is the same as the organisms in the out-group, then that character state can be considered plesimorphic [24]. The outside taxon is called the out-group and the organisms being classified are the in-group. The only way a homologous feature could be present in both an in-group and an out-group, would be for it to have been inherited by both groups from an ancestor older than the ancestor of just the in-group.

In summary, the application of cladistics to manufacturing systems has opened up a new and more holistic way to classify and link evolutionary changes to main manufacturing changes of the past and present. ‘Manufacturing’ cladistics has now been successfully applied to discrete manufacturing systems [2] [3] [4] [22] [29] [31] [32] and to aerospace supply chains [40] [41] but, to this point in time, is sector specific (e.g., the automobile and hand-tool industries). The aim here is to develop a general classification system that spans sectors.

2 METHODOLOGY

Constructing the cladistic classification is often viewed as an eight-step sequential process (see below) following what is known as the ‘waterfall’ method [22]. However, in practice this fails to represent the actual process. The first six steps essentially overlap, are often concurrent activities and help refine each of the other steps’ outcomes.

Problem Definition: The classification process should begin by stating clearly the nature of the problem to be solved which provides the basis to understand the relation between the Species and the characters that define the Species [20] [33]. Also, it is necessary for setting the boundaries for the classification to avoid ambiguity in later steps of the research [30].

Determine the Clade (Taxon): A clade is a taxon that in this case consists of the manufacturing systems under study, along with their common and most recent ancestors. The ancestor and all the other Species must be identified manually. This is an iterative task where the Species and definition of each Species can be modified by the researcher [22].

Determine the Characters: A character is any variable, feature or attribute, which forms the basis for classificatory significance [22]. Taxonomic characters perform two functions: firstly, they have a diagnostic aspect uniquely specifying a given taxon and an emphasis on the differentiating properties of taxa is particularly strong at the level of lower categories [32]; and, secondly, they function as indicators of relationships; a property that makes them especially useful in the study of the higher taxa [24]. The character selection is a manual

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

process involving secondary data (i.e., literature, company records, annual reports, business plans and technical data such as layout plans, control/scheduling strategies, etc). Candidate characters are considered and then rejected if they are irrelevant or if they create ‘noise’ in the data table [22] [32]. Qualitative methods, such as observations, field visits, and interviews and discussions with key personnel, are also typically implemented.

Code the Characters: Numbering characters, helps with both ordering [24] and making decisions concerning whether they exist in the entities under study [22]. Different states for each character are proposed and an indication given of what Species possess them [39].

Ascertain Character Polarity: Character polarity is defined by the distinction between a primitive character or state which is present in ancestral Species and a derived character or state which is not present in the ancestor [22].

Estimating Phylogeny and Constructing the Conceptual Cladogram: Cladograms are constructed by grouping Species that share a common root and evolutionary history. Cladograms may be constructed manually or through dedicated software such as MacClade: Analysis of Phylogeny and Character Evolution [27], used in this research, which offer manipulation tools, in which characters and states can be ordered, weighted and traced.

Construction of the Factual Cladogram: This step largely involves contemporary organisations, i.e., Specimens of Species [31], and is more quantitative in nature, i.e., surveying a representative sample of Species and the Specimens within. The aim is to test the hypotheses inherent in the conceptual cladogram. Any conflicts are then resolved leading to a full factual cladogram [32]. Emphasis on continuous hypothesis testing is the strength of cladistics as an instrument for theory development and advancement [30] [33] [37].

Decide Taxa Nomenclature: This stage deals with the labelling or naming of the manufacturing systems. This should conform to the principles of biological nomenclature. In short, names should convey the essence of the entity (typically their main character), be unambiguous, and ensure universal communication [32]. In the Linnaean tradition, Species are given a binominal or two-term name relating to the Species and the Genus it belongs to.

3 RESULTS AND DISCUSSION

For step 1 (above), the classification problem – the Species – is defined as follows:

‘A coherent set of processes which, depending on the complexity of that being manufactured, represents a significant stage in production and produces a coherent, single or family of parts, components, modules or final products. The boundary is not necessarily a whole factory system, which can be set out in modular fashion and contain plant within plants (in effect an ecology of different species), but individual workstations, cells or plants, the latter being a relatively small set of workstations or cells.'

The following relates to work undertaken during steps 2-6. Therefore, using the cladistic approach, the evolutionary relationships between forty-four candidate Species of

manufacturing systems, using ‘descriptors’ drawn from a library of twelve characters with a total of sixty-six states (see Table 1), are hypothesised, described and presented

diagrammatically (see

Figure 1). The manufacturing Species are then organised in a hierarchical classification with thirteen Genera, six Families and three Orders under one ‘Class’ of discrete manufacturing

(see

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

Figure 2).

Table 1. Primary characters and states of manufacturing systems

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

Figure 1. Cladistic classification of discrete manufacturing systems

Figure 2. Hierarchical (Linnaean) classification of discrete manufacturing systems

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

3.1 An evolutionary account

The evolutionary history, depicted in

Figure 1 must begin with an Out-group [11] [22] [24], which represents Self Production. This primitive system of manufacturing shares many of the characters to the in-group or clade passed on from a common ancestor. Self Production manufactures articles for personal use, in a fixed position (CS 1-1), in one site (CS 2-1) and usually in or around the place of living. Simple, universal, processing techniques and tools (CS3-1) are employed, in the form of manual or hand tool manipulation (CS 6-1). All the necessary processes are performed and the full article produced, by the one person (CS4-1) in one go, i.e., without WIP or ‘buffer’ between the processes (CS 5-1). All material handling (i.e., both primary (between processes) and secondary (within processes) is primarily manual (CS 7-1; CS 8-1) and, in some instances, mechanised (primitive pulleys, winches, etc.).

The first Species to evolve from the common ancestor starting what is now the Class of Discrete Manufacturing is the Product Centred Workshop [1] and belongs to the Multi-Family Order of manufacturing systems. In this Order, variations of the above-mentioned Characters are evident (with the exception of material handling) in addition to two new characters that emerge – the style of management and the power over resources that are managed in project-managed products. The most significant CS change is the General Layout Approach with the fixed position layout (CS 1-1) being the most defining CS for the Fixed-Position Family and the process layout (CS 1-2) the most defining CS for the Process Family [43]. The Fixed-Position Family comprises two Genera, the Product Centred and the Project. The first Species of the Product Centred Genus is the Product Centred Workshop [43], whose the primary difference from the Out-group is that an entrepreneurial spirit (CS 9-1) has emerged in which the products made are sold to customers. The second Species is the Product Centred Yard [12]. Here, a variation of the Management Style Character is evident. In the Product Centred Yard, products are more complex, require more workers, who still perform significant product processes, but only produce part of the product (CS 4-2) albeit a significant part. A more project-managed (CS 9-2) environment is required in order for the project manager (PM) to get the best out of the project resource pool (CS 10-1).

Through a common ancestor of the Product Centred Yard, the Project Genus [17] [34] evolves and represents variations in the location of production, management style and PM resource power [42]. The most defining CS for this Genus is with the production taking place at a remote location (CS 2-2) where all resources are brought to a specific one-off location, i.e., the customer. Two types of project systems have evolved. The first includes the Project Pure and the Project Virtual Species [8]. The former is differentiated from the Product Centred Yard purely by the production taking place at a remote location (CS 2-2). The latter represents a variation in the size and nature of the resource pool – the Product Centred Yard and Project Pure have an intra-organisational resource pool (CS 10-1) (within one organisation but without any functional divisions), whereas the Project Virtual, which spans different organisations (i.e., different companies), has an inter-organisational resource pool (CS 10-4). The other type of project system has evolved within functionally organised companies, in which projects need to be implemented and managed [43]. One Species is called the Project Functional and represents those systems in which one-off products are produced within a certain function (typically the engineering division). Here the functional manager is the project manager (CS 10-2) and has responsibility over both the function (department or division) and the project. The Project Matrix [8], represents projects in which cross-functional resources are needed, requiring a specialised project manager who has power to second specific functional resources (CS 10-3). The Project Agile [28] [34] seems to be a recent addition to the Genus where agile project management techniques (CS 9-3) are

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

employed in response to the inflexibility associated with formal project management techniques.

The second Family of the Multi-Family Order has been named the Process Family due to a variation of the general layout approach character. The first Species in this Family, the Neocraft Shop [31], which is one of two Species of the Genus, evolved from a common ancestor of the Product Centred Workshop and shares the majority of CSs except the implementation of a process layout (CS 1-2), where the product moves to each machine or process. This change was primarily due to the nature (size and weight) of the mechanised machines that were introduced, i.e., manually operated machines (CS6-2), where the machines would be placed and fixed in certain areas. The second Species, the Neocraft Jobshop [5], emerging with the industrial revolution, built on the early evolutionary success of the Neocraft Shop with a scaling up of capability, i.e., with additions of more of the same key machines and processes (CS 3-2). With these systems, more workers are employed who concentrate on their process expertise and perform significant product processes but only produce part of the product (CS 4-2). This type of system however creates fairly significant WIP between processes (CS 5-2).

The Scale Genus takes this further with the large-scale additions of most or all machines (CS 3-3). The first of the Species in the Genus, the Scale Batchshop, making an appearance not long after the Neocraft Jobshop, exhibits two other fundamental changes, which differentiates the Species. With the first change, although significantly more workers are required to meet the much larger orders there is a lack of skills available in the marketplace which leads to the employment of unskilled workers and to a deskilling in the production process where workers perform single or a very limited set of processes (CS 4-3). With more workers, a requirement for a change in management style leading to a more centralised approach (CS 9-4). The next two Species in the Genus, the Scale Linked Batch [17] and the Scale Nagare [19], lead to a major bifurcation in the evolutionary scheme with the ancestor of the former leading to the Product Line Order around the turn of last century, and the ancestor of the latter making way for the Group Technology Order; both Species still have the process layout but some machines are ‘virtually’ linked either to form a product line (CS 1-2/31) [17] as with the Scale Linked Batch (which is the only fundamental change in terms of the primary characters) or to form a cell (CS 1-2/42) representing the Scale Nagare. For the Scale Linked Batch, the layout is predominantly process based, but certain machines in each area are dedicated fully to one particular product [17]. The Scale Nagare also implements some Lean principles of both multi-skilling workers where they take responsibility for all product family processes, and of removing waste in the form of in-process buffer (CS 5-4).

The first Species of the Product Line Order, evolving from the ancestor of the Scale Linked Batch, is the Unpaced Asynchronous, which belongs to one of two Genera under the Family of Manual. The Unpaced Asynchronous can be distinguished from the Scale Linked Batch by the implementation of single dedicated machine/process types (CS 3-4) arranged in a product layout (CS 1-3). The Unpaced Synchronous is the second Species in the Genus and is associated with assembly lines [44], where processes are done manually or using hand-tools (CS 6-1), which creates the opportunity of balancing the line to minimise in-process buffers (CS 5-3). The Machine Paced Genera introduces an automated primary material handling system (PMHS) [6]. Furthermore, the only differentiating factor within this Genus is the exhibition and exploration of different types of PMHS [31]. The Machine Paced Stop & Go has an intermittent PMHS where the conveyor, in-line cart, etc., stops for every

1 This Species exhibits two states of the character ‘general layout approach’; both the process layout (CS 1-2) and the product layout (CS 1-3). 2 This Species exhibits two states of the character ‘general layout approach’; both the process layout (CS 1-2) and the group technology layout (CS 1-4).

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

process/workstation (CS 11-1); the product is usually of moderate size and includes, for example, over 30 workstations; processing time at each workstation is between say 30 seconds and several minutes (Takt timed). The Machine Paced Continuous [12] features a continuous PMHS and the operator performs the processes whilst the product/part is being carried by the PMHS (CS 11-2); operator process times are very quick, typically 1-10 seconds. The Machine Paced Pick & Drop also exhibits a continuous PMHS but in this instance the operator removes the part/product from the conveyor to perform process(es) then returns it (CS 11-3); process times here are between, for example, 10-30 seconds. Specimens can be found, for example, in the final packaging line. The Machine Paced Comb & Spine differs slightly as operator removes part/product from the conveyor to perform process(es) but feeds it to another conveyor (CS 11-4). This is typical of both the mixed model and postponement strategies. Both the Machine Paced Moving and the Machine Paced Sliding Station also have a continuous PMHS but whereas with the former the operators perform process(es) by ‘walking/moving’ with the in-line cart (CS 11-5), the latter has some workstations that ‘slide’ past other workstations to perform processes (CS 11-6).

The second and final Family, the Automated [17], of the Product Line Order, replaces human processing with machine processing. There are two Genera, the Transfer and the Robot with variations in the characters of process technology type and automated PMHS type explored. The Transfer Intermittent, like the Machine Paced Stop & Go, features an intermittent PMHS where the conveyor, in-line cart, etc., stops for every process (CS 11-1), but also introduces automated machines that are non-CNC (CS 6-4) and combines the secondary material handling with the PMHS (CS 8-2). The role of the operator shifts to overseeing and monitoring processes (CS 4-4). The second and final Species in the Genus, is the Transfer Continuous [13]. Here the Automated PMHS type changes to a continuous cycle where parts/products are automatically processed whilst in motion. Within this Genus, the Robot Unidirectional [17], as the name suggests, differs by way of the introduction of robots (CS 6-6) as the primary process technology type. The Robot Cyclic, typically used with products where a simple sequence of fairly straightforward welding or machining processes take place. Here the PMHS stops for every process but cycles around where the product/part is firstly removed from the pallet and another is added for processing (CS 11-8).

The third Order of Group Technology evolves from the ancestor of the Scale Nagare, primarily with the introduction of a group technology [17] layout (CS 1-4) which all Species share. An additional character also appears, that of cell buffer. Although the Species in this Group Technology Order really proliferated during the 1960-70’s onwards. The Lean Family is composed of two Genera – the Cell and the U-Line, both of which explore variations in operator task types and responsibilities, process technology types and cell buffer types [49]. The Cell Chase, the first of four Species in this Genus [18], introduces buffer at the level of the cell which decouples the cells creating independent cells (CS 12-1); the operator task type/responsibility also changes from performing single or a very limited set of processes to performing all part family processes (CS 4-6) and chases the part through the cell hence the name. The Cell Agile, differs from the Cell Chase [35] through the exploration of modular mechanised machine tools (CS 6-3) which are typically on wheels/casters and can be quickly re-configured in response to changes in demand. The Cell Zonal [35], also evolving from a common ancestor of the Cell Chase, introduces two or more operators whom share cell processes in zones (CS 4-7). The Cell Split takes this further but introduces three or more operators processing a part each which are then brought together for final processing or assembly (CS 4-8). The U-Line [35] Genus comprises three Species, the first of which is the U-Line Decoupled. For this the buffer between cells (CS 12-2) is eliminated resulting in a fully integrated set of cells in the shape of a large U-Line. With the introduction of modular mechanised machines (CS 6-3), the U-Line LeAgile is created, which can be quickly re-

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

configured to meet changing demand patterns. The U-Line Multi explores cell buffer further and eliminates buffer at the line level (CS 12-3) creating fully integrated U-Lines.

There are three Genera under the FMS Family – the Semi-Flexible, the Flexible and the Robotic which are defined through their exploration of variants of process technology types, automated PMHS types and secondary material handling systems [47]. All Species share the following two features: an automated PMHS (CS 7-2), and operators that solely programme, oversee and monitor processes (CS 4-9). The Semi-Flexible Genus, most of which share the feature of the secondary material handling combined with the PMHS (CS 8-2), begins with the Semi-Flexible Bypass which introduces CNC machine tools (CS 6-5) and an intermittent unidirectional PMHS that can bypass processes as required (CS 11-9). The Semi-Flexible Desktop [46] and Semi-Flexible Square Foot [49] further explore different process technology types in the form of micro machining units (MMUs) and modular MMUs, respectively. The common ancestor of the Semi-Flexible Bypass also leads to the Semi-Flexible Rotary Indexer [14] and Semi-Flexible Bidirectional Self Feed through variants in automated bidirectional PMHS types with the former using a rotary indexing PMHS (CS 11-10) and the latter using conveyors, or something similar, that move in two directions (CS 11-11). The latter also moves to automated secondary material handling (CS 8-3), a feature that is shared with the Flexible and Robotic Genera. All but the Species in the Flexible [17] Genus can also be defined by the ‘mobile’ variants of the automated PMHS types. For the Flexible Ladder, which is laid out in the shape of a ladder, the evolutionary distinguishing feature is the use of automated guided vehicles or AGVs [38] (CS 11-12). The Flexible Open Field, where there is no specific layout, self-guided vehicles (SGVs) are typically used (CS 11-13). The Flexible Reconfigurable [7] introduces artificially intelligent (AI) SGVs (CS 11-14) in addition to modular CNC machine tools (CS 6-7). The Flexible Holonic [10] [21] introduces autonomous CNC machine tools, i.e., with AI (CS 6-8). The Flexible Robot Centred has a robot (CS 11-15) as the PMHS, which is also a feature of the Robotic [36] Genus. All Species have robots as the main process technology type as with The Robotic Cell (CS 6-6). The Robotic Plug & Produce introduces modular robots (CS 6-9) for reconfigurabilty. The final Robotic Adaptive exhibits autonomy and adaptability through robots with AI (CS 6-10).

4 APPLICATIONS

This paper is also concerned with the useful application of this work and the development of a benchmarking tool so that manufacturing organisations in need of change can easily locate where they are in an evolutionary landscape, identify where they want to be, and how to get there. The method suggested is characterised by a speed-read technique of quick Species identification. The Linnaean hierarchy is the map into which the firm can search out its closest present manufacturing system (see Figure 3). The end-user interrogation begins by asking relatively few questions designed to quickly narrow down to the manufacturer’s production system using the hierarchy. At each of the main branches on the hierarchy is a ‘primary evolutionary character’. The system is able to quickly identify a general production system with only 4 questions, illustrated here through the C1 case study manufacturer.

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

Figure 3. Speed read process using the hierarchical classification

In the case of C1, is: Question 1: which best describes the relationship between the product and the manufacturing system? For C1 the answer is ‘the system is dedicated to a single product or a limited number of different variants of the product. Machines and processes are in a dedicated sequence for one product or a limited number of models’. The state being Character 1 (General Layout Approach) and State 3 (Product Layout). The result is that C1 belongs to the Line Order. Next, C1 is directed to Question 7: ‘what best describes the process technology type?’ C1 answers ‘mechanised machines’, i.e., CS 6-2. The result is the Manual Line Family. Next, C1 is directed to Question 8: ‘what best describes the primary material handling system?’ C1 answers ‘manual/mechanised’, i.e., CS 7-1. The result is the Unpaced Genus. Next, C1 is directed to Question 9: ‘what best describes the in-line buffer?’ C1 answers ‘buffer between workstations’, i.e., CS 5-2. The final result is Species 13: Asynchronous Unpaced, Manual, Product Line.

The gap between the present manufacturing performance of a firm and how it wishes to perform has to be addressed focusing on three things: a) assess the current performance features, b) derive a set of target levels of performance, and c) compare the present against the target performance. This involves a comparison between the general product, process and system characteristics of the end-user ‘variant’ Specimen and the ‘ideal’ or ‘textbook’ Species. This knowledgebase draws from a library of our 82 characters and their states ordered into 14 groups of CS sets. The diagnostic visualisation tool applied is a Kiveat diagram (or radar plot, spidergram, etc.). Through a comparison of the C1 case study (Specimen) and the ideal Species, a list of recommendations can be given. Five of the fourteen groups are given for illustration below:

1. General Product characters:- Fitness score: 83.3%. Diagnosis: five of the six states are identical for the current and ideal cases. C1 has medium order volume, while the ideal is high. Recommendations: Consider increasing order volume.

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

2. Product competitive priorities:- Fitness score: 50%. Diagnosis: two of the four states are identical for the current and textbook cases. C1 has a high priority on performance design, while the ideal is low. Recommendations: Consider redcuing priority on performance design. Diagnosis: C1 have medium priority on product cost, while the ideal is high. Recommendations: Consider reducing product costs.

3. Process technology:- Fitness score: 66.7%. Diagnosis: Two of the three states are identical for the current and ideal/typical cases. C1 have limited universal-machine/process duplication, while ideal is single dedicated-machine/process types. Recommendations: Consider having single, dedicated machine/process types.

4. General processes:- Fitness score: 40%. Diagnosis: two of the five states are identical for the current and textbook cases. C1 have moderate process times, while the ideal is short. Recommendations: Consider reducing process times. Diagnosis: For C1 bottlenecks are the key process, while the typical is none. Recommendations: Consider removing bottlenecks. Diagnosis: C1 have variable costs on setups, while typical is high. Consequence: A consequence of single dedicated machines are high setup costs but as they are rarely re-set this should not be a problem.

5. Production competitive priorities:- Fitness score: 71.4%. Diagnosis: five of the seven states are identical for the current and ideal/typical states. C1 places a medium priority on competing on production cost, while the ideal is high priority. Recommendations: Consider reducing production costs. Diagnosis: C1 places medium priority on volume flexibility; the ideal is low. Consequence: MTS and high volume typically results in low volume flexibility.

In summary, the overall fitness of C1 compared to the textbook Asynchronous Unpaced, Manual, Product Line is 50%. The main challenge is with Product Competitive Priorities and General Processes. If these are corrected they would have a fitness score of 76%, a significant improvement. However, this is only a preliminary score as the system of fitness scoring needs to be refined. Some characters and states need weighting for significance and will outweigh others. The distinction between recommendations and consequences also needs further investigation as well as the knock on effects. For example, single dedicated machines have multiple impacts on: setups; flexibility; capacity; investment; and, overheads.

5 CONCLUSIONS

The aim of this research was to unify the various discrete manufacturing system classifications, typologies and taxonomies in the literature. Two complementary conceptual schemes were developed – hierarchical and cladistic classifications. After defining the classification problem and identifying the range of ‘Species’ to include, the evolutionary-unique, product-, process- and system characteristics of each Species were determined, systematically coded, and subjected to polarity ascertainment. This in turn led to the conceptual cladogram. Forty-four candidate manufacturing systems, using ‘descriptors’ drawn from a library of twelve characters and sixty-six states, are described and presented diagrammatically. The complementary classification organises this information hierarchically and groups Species under ‘Genera’, ‘Families’ and ‘Orders’ based on evolutionary proximity. The results presented here are largely conceptual in nature and based on secondary data including the Operations and Production Management literature, company records, annual reports, business plans and technical data (layout plans, control/scheduling strategies, etc). The classifications are, however, arguably and demonstrably consistent with, and synthesise many of the established typologies in the literature. The next stage of research, the testing and refinement of theory, which produces the final factual cladogram, will be grounded in quantitative data from a representative sample of discrete manufacturers using a web-based

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

measurement instrument. Applications, demonstrated by the C1 use case, include: speed-read diagnostics, detailed review of characteristics, a detailed profile, and solution landscape for recommendations. This web-based expert system and diagnostic tool, a ‘cladistic driven general requirement engineering tool’ complements a larger software architecture in the COPERNICO project (www.copernico.co). The aim is to simplify and make accessible essential tools for rapid design, simulation and virtual prototyping of factories.

REFERENCES

[1] Alizon, F.; Dallery, Y.; Essafi, I.; Feillet, D. (2009): Optimising material handling costs in an assembly workshop, International Journal of Production Research, 47(14), pp. 3853-3866.

[2] Baldwin, J.S.; Allen, P.M.; Winder, B.; Ridgway, K. (2003): Simulating the Cladistic Evolution of Manufacturing, Innovation: Management, Policy and Practice, 5(2-3), pp. 144-156.

[3] Baldwin, J.S.; Allen, P.M.; Winder, B.; Ridgway, K. (2004): Modelling Manufacturing Evolution: Thoughts of Sustainable Industrial Development, Journal of Cleaner Production, 13(9), pp. 887-902.

[4] Baldwin, J. S.; Allen, P. M.; Ridgway, K. (2010): An Evolutionary Complex Systems Decision-Support Tool for the Management of Operations, International Journal of Operations and Production Management, 30(7), pp. 700-720.

[5] Betrand, J. W. M.; van Ooijen, H. P. G. (2008): Optimal work order release for make-to-order job shops with customer order lead-time costs, tardiness costs and work-in-process costs, International Journal of Production Economics, 116(2), pp. 233-241.

[6] Boysen, N.; Fliedner, M.; Scholl, A. (2007): A classification of assembly line balancing problems, European Journal of Operational Research, 183(2), pp. 674-693.

[7] Bruccoleri, M.; Pasek, Z. J.; Koren, Y. (2006): Operation management in reconfigurable manufacturing systems: Reconfiguration for error handling, International Journal of Production Economics, 100(1), pp. 87-100.

[8] Burke, R. (2003): Project management: planning and control techniques, John Wiley & Sons, Chichester.

[9] Camin, J.H.; Sokal, R.R. (1965): A Method for Deducing Branching Sequences in Phylogeny, Evolution, 19(2), pp. 311-326.

[10] Cheung, H.M.E.; Yeung, W.H.R.; Ng, H.C.A.; Fung, S.T.R. (2000): HSCF: a holonic shop floor control framework for flexible manufacturing system, International Journal of Computer Integrated Manufacturing, 13(2), pp. 121-138.

[11] Darwin, C. (1859): The Origin of Species, Oxford World’s Classics paperback 1998, New York.

[12] De Toni, A.; Panizzolo, R. (1992): Repetitive and intermittent manufacturing: comparison of characteristics, Integrated Manufacturing Systems, 3(1), pp. 23-37.

[13] Dhoib, K.; Gharbe, A.; Landolsi, N. (2009): Throughput assessment of mixed-model flexible transfer lines with unreliable machines, International Journal of Production Economics, 122(2), pp. 619-627.

[14] Dolgui, A.; Guschinsky, N. N.; Levin, G. M. (2009): Graph approach for optimal design of transfer machine with rotary table, International Journal of Production Research, 47(2), pp. 321-341.

[15] Dollo, L. (1893): Les Lois de l’evolution, ulletin de la Société Belge de Géologié, de Paléontologie, 7(2), pp. 164-166.

[16] Hennig, W. (1950): Grundzuge einer Teorie der phylogenetischen Sytematic, Deutsche Zentralverlag, Berlin.

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

[17] Hill, T. (2005): Operations Management, 2nd Edition, Basingstoke, England: Palgrave MacMillan.

[18] Jajodia, S.; Minis, I.; Harhalakis, G.; Proth, J.M. (1992): Computerised layout solution using simulated annealing, International Journal of Production Research, 30(1), pp. 95-108.

[19] Kannan, V. R.; Ghosh, S. (1996): A virtual cellular manufacturing approach to batch production, Decision Sciences, 27(3), pp. 519-539.

[20] Kluge, A.G.; Farris, J.S. (1969): Qualitative phyletics and the evolution of anurans, Systematic Zoology, 18(1), pp. 1-32.

[21] Lee, F. D.; Banjeree, S. (2010): Design and dynamic analysis of an arch-type desktop reconfigurable machine, International Journal of Machine Tools and Manufacture, 50(6), pp. 575-584.

[22] Leseure, M.J. (2000): Manufacturing strategies in the hand-tool industry, International Journal of Operations and Production Management, 20(12), pp. 1475-1487.

[23] Linnaeus, C. (1958): Systema Naturae (editio decimal, reformata). Tom.I. Laurentii Salvii, Holmiaa.

[24] Lipscomb, D. (1998): Basics of Cladistic Analysis, Weintraub Program in Systematcis and Department of Biological Sciences, George Washington University, Washinton D.C.

[25] Lorenz, K.Z. (1974): Analogy as a source of knowledge, Science, 185(3), pp. 229-234. [26] MacCarthy, B.L.; Fernandes, F.C.F. (2000): A multi-dimensional classification of

production systems for the design and selection of production planning and control systems, Production Planning and Control, 11(5), pp. 481-496.

[27] Maddison, D. R.; Maddison, W. P. (2003): MacClade 4: Analysis of Phylogeny and Character Evolution (V.4.06). Sinauer Associates, Inc. Sunderland, Massachusetts, USA.

[28] Mafakheri, F.; Nasiri, F.; Mousavi, M. (2008): Project agility assessment: an integrated decision analysis approach, Production Planning and Control, 19(6), pp. 567-576.

[29] McCarthy, I. (1995): Manufacturing classification: Lessons from organizational systematics and biological taxonomy, Integrated Manufacturing Systems, 6(6), pp. 37-48.

[30] McCarthy, I.P. (2005): Toward a phylogenetic reconstruction of organizational life, Journal of Bioeconomics, 7(3), pp. 271-307.

[31] McCarthy, I.; Leseure, M.; Ridgway, K.; Fieller, N. (1997): Building a manufacturing cladogram, International Journal of Technology Management, 13(3), pp. 269-286.

[32] McCarthy, I.; Ridgway, K. (2000): Cladistics: a taxonomy for manufacturing organizations, Integrated Manufacturing Systems, 11(1), pp. 16-29.

[33] McKelvey, B. (1982): Organizational Systematics – Taxonomy, Evolution, Classification, University of California Press, Berkely, Los Angeles, London.

[34] Meade, L. M.; Sarkis, J. (1999): Analyzing organizational project alternatives for agile manufacturing processes: an analytical network approach, International Journal of Production Research, 37(2), pp. 241-261.

[35] Miltenburg, J. (2001): U-shaped production lines: A review of theory and practice, International Journal of Production Economics, 70(3), pp. 201-214.

[36] Mohamed, A. M.; Geismar, H. N.; Sethi, S. P.; Sriskandarajah, C. (2005): Sequencing and scheduling in robotic cells: recent developments, Journal of Scheduling, 8(5), pp. 387-426.

[37] Quicke, D.L.J. (1993): Principles and techniques of contemporary taxonomy, Blackie Academic & Professional, London.

[38] Rajotia, S.; Shanker, K.; Batra, J.L. (1998): An heuristic for configuring a mixed uni/bidirectional flow path for an AGV system, International Journal of Production Research, 36(7), pp. 1779-1799.

[39] Ridley, M. (1993): Evolution, Blackwell Scientific Publishers.

MITIP 2012 Modern Information Technology in the Innovation Processes of Industrial Enterprises

[40] Rose-Anderssen, C.; Baldwin, J.S.; Ridgway, K.; Allen, P.M.; Varga, L.; Strathern, M. (2009): A cladistic classification of commercial aerospace supply chain evolution, Journal of Manufacturing Technology Management, 20(2), pp. 235-257.

[41] Rose-Anderssen, C.; Baldwin, J.S.; Ridgway, K. (2011): Commercial aerospace supply chains: the empirical validation of an evolutionary classification scheme, Journal of Manufacturing Technology Management, 22(1), pp. 66-89.

[42] Shenhar, A.J. (1998): From theory to practice: toward a typology of project-management styles, IEEE Transaction on Engineering Management, 45(1), pp. 33-48.

[43] Slack, N.; Chambers, S.; Johnston, R.; Bates, A. (2006): Operations and process management – Principles and practice for strategic impact, Prentice Hall – Financial Times.

[44] Smunt, T.L.; Perkins, W.C. (1985): Stochastic Unpaced line design: review and further experimental results, Journal of Operations Management, 5(3), pp. 351-373.

[45] Sober, E. (1983): Parsimony in systematics: philosophical issues, Annual Review of Ecology and Systematics, 14(4), pp. 335-357.

[46] Son, H.; Choi, H.-J.; Park, H. W. (2010): Design and dynamic analysis of an arch-type desktop reconfigurable machine, International Journal of Machine Tools & Manufacture, 50(6), pp. 575-584.

[47] Tacquard, C.; Martineau, P. (2001): Automatic notation of the physical structure of a flexible manufacturing system, International Journal of Production Economics, 74(1-3), pp. 279-292.

[48] Wulfsberg, J.P.; Redlich, T.; Kohs, P. (2010): Square Foot Manufacturing: a new production concept for micro manufacture, Production Management, 4(1), pp. 75-83.

[49] Womack, J.; Jones, D.; Roos, D. (1990): The Machine that changed the World, Rawson Associates, New York.

ACKNOWLEDGEMENTS

The authors acknowledge the support of the European Commission through the 7th Framework Programme under NMP-2008-3.4-1 Rapid Design and Virtual Prototyping of Factories (COPERNICO: contract number 229025-2).