User:Noble Krypton

Active Knowledge Modeling (AKM) AKM is capturing knowledge involved in building workplaces, in supporting work execution and knowledge generated by work execution. AKM was co-founded by Frank Lillehagen (who between 1990 to 1993 discovered the foundation of AKM), Dag R. Karlsen and Håvard D. Jørgensen.

Overview
Modeling should be as natural as drawing, sketching, and scribbling and should provide powerful services to capture work-centric, work-supporting and generative knowledge, for preserving context and ensuring reuse. The development of AKM is the solution to the aforementioned issue of modeling. Although AKM has the potential value and usage across a large range of knowledge creation and knowledge representation tasks, the main focus is on the use of these techniques in providing IT support for performing creative and innovative work. The AKM technology is about discovering, externalizing, expressing, representing, sharing, exploring, configuring, activating, growing and managing enterprise knowledge. An AKM solution is about exploiting the web as a knowledge-engineering medium, developing knowledge model-based families of platforms, model-configured workplaces and services.

The main objective of AKM technology is to make explicit and exploit knowledge that add value to the enterprise and can be shared by business services and users for improving the agility and the performance of the enterprise and for getting much more value from IT and Web technologies

The AKM approach and integrated methodologies captured in Active Knowledge Architecture (which happens to be a project or sector-specific knowledge landscape built using sector and project specific Information-Role-Task-View (IRTV) languages to capture contents from customer) enables us to capture, share and benefit from situated, work - generative knowledge that otherwise would have remain tacit in the minds of those involved.

History
The industrial needs and thinking that sparked the initial development of the AKM technology in the late 1980's was inspired and influenced by Industrial War Rooms (Figure 2.1)

Fig 2.1: The industrial war room-inspired AKM thinking War rooms were meeting places to discuss the many known, but not all described and considered, and often forgotten, dependencies and relationships between objects, structures, views, and responsible people. The core concepts of AKM such as the knowledge spaces were conceived at Volvo cars around 1990. The ten founding discoveries of the AKM technology date back to 1990, when a team headed by Frank Lillehagan, was engaged in innovative car projects with Volvo cars, Gothenburg. There discoveries made in the early 1990's are in current language described as follows:  Enterprise knowledge exists in nested multidimensional bounded spaces, and in delimited layers and domains. The spaces and dimensions involved in enterprising and product design are neither linear nor orthogonal; they reflect human mental models with perspective views. Most enterprise aspects and views are mutually inclusive – as a consequence of the nested knowledge spaces, ref., the war-room thinking (Zakis 2007). The world is both perspective view (method) and object-oriented – we need to integrate mental and object-oriented computer models and views, producing work-sensitive models. Mental models in the human brain, with perspective and computed views, content and context contributed by specific roles, are poorly understood and currently not exploited by IT people. Existing models are based on and bounded by diagrams, charts, and mathematical formalisms; there is no capture of situated knowledge,exploiting the intrinsic knowledge properties. Process- and work-flow, and time-dimension phase dependencies must be relaxed or expanded, and minimized by providing intelligent working environments. Present Software Engineering approaches will never adequately handle properties, parameter trees, and multiple value sets. Learning, design, and problem-solving are intimately related, use similar methods, and have similar service and viewing needs. Deployed legacy systems are a challenge, but small compared to the prevailing legacy thinking. Vaults of information documents describing design rules, materials, reference models, and more should become sharable active knowledge.  The discoveries imply that knowledge exists both as object-oriented IT structures and as perspective method-oriented structures, such as the lifecycle view, where different roles perform tasks to add methods, content, and context in some common views as well as in role-specific views. Thus the foundations of the AKM technology were discovered in industrial innovation projects, attempting to build digital product models according to the war room thinking.

Scientific Foundations of AKM Technology
AKM was discovered based on some scientific methods and theories that have already been used in some modeling languages such as conceptual modeling language.

The scientific methods and theories that provide explanations and principles for the AKM discoveries as illustrated in Figure 3.1 are:


 * Epistemology
 * Human learning, pedagogy and psychology
 * Natural language, linguistics and semiotics
 * Process design and Engineering
 * Organizational development and learning
 * Product design and Engineering
 * Systems Engineering

Fig 3.1: Scientific methods and theories that provide explanations and principles for the AKM discoveries

Epistemology
Organizational change may be viewed from different philosophical points of view. Two common sets of assumptions are the objectivistic belief system and the constructivistic belief system. These sets of assumptions may be distinguished through differences in ontology (what exists that can be known), epistemology (what relationship is there between the knower and the known) and methodology (what are the ways of achieving knowledge)

Organizations are realities constructed socially through the joint actions of the social actors in the organization as illustrated in the figure 3.2. An organization consist of individuals who view the world in their own specific way, because each of them has different experiences arisisng from work and other areas. The local reality refers to the way an individual perceives the world in which he/she acts. When the social actors of an organization act they externalize their local reality. In constructing organizational reality, social actors make something that other actors have to relate to by being part of the organization. Internalization is the process of making sense out of the actors, institutions, artifacts, etc in the organization and making this organization reality part of the individual local reality.

Fig 3.2: Social construction in an organization

Human learning, pedagogy and psychology
Activity theory happens to be an important area in human learning, pedagogy and psychology. According to activity theory, human knowledge, learning,and activity in general are fundamentally related to collective systems engaging in goal-directed action on the basis of underlying motives shared among the activity participants. Activity theory stresses that human activity has to be regarded in a holistic manner. The basic structure of human activity is illustrated in the figure 3.3.

Fig 3.3: The basic structure of human activity

From the figure 3.3 it can be deduced that connection between the subject and the object is necessitated by an instrument (tools) or instrument artifacts that facilitate processes of work.Subjects change and develop themselves in fulfilling the activity, which means purposefully changing natural and social reality. The activity object is a motive and a driving force. Knowledge is fundamentally related to practices: “Knowing is defined only in the context of specific practices, where it arises out of the combination of a regime of competence and an experience of meaning” Wenger (1999) .The distinction between explicit and tacit knowledge follows from Polanyi (1966) : Explicit or codified knowledge is transmittable in a formal systematic language, while tacit knowledge has a personal quality, which makes it hard to formalize and communicate. Nonaka and Takeuchi (1995) came out with four patterns of interaction between tacit and explicit knowledge commonly called modes of knowledge conversion as illustrated in Fig. 3.4.

Fig 3.4: Modes of knowledge conversion

The internalization mode of knowledge creation is closely related to “learning by doing,” hence action is deeply related to the internalization process. When tacit and explicit knowledge interacts, innovation emerges. Nonaka and Takeuchi propose that the interaction is shaped by shifts between modes of knowledge conversion, induced by several triggers as illustrated in Fig. 3.4, we have the socialization mode starting with building a field of interaction facilitating the sharing of experience and mental models. This triggers the externalization mode by meaningful dialogue and collective reflection where the use of metaphor or analogy helps articulate tacit knowledge, which is otherwise hard to communicate. The combination mode is triggered by networking newly created knowledge with existing organizational knowledge, and finally learning by doing triggers internalization.

Natural language, linguistics and semiotics
Natural language is essential for industrial nomenclature and logistics. Words are the tools by means of which individuals grasp the thoughts of others and with which they do much of their own thinking. They are the "tools of thought" and may not express enough situated knowledge and context to be useful of their own. Instead of words, models can be used as a primary knowledge representation approach. The field of semiotics is about the science of signs and what they refer to. According to FRISCO report (Falkenberg et al. 1996), the means of communication and related areas can be examined in a semiotic framework. The below semiotic layers for communication are distinguished,forming what they term a semiotic ladder. Model denotations are signs,and thus they have considered the semiotics of models.The key concepts to be included in information systems models is regarded to be:
 * Physical: use of various media for modeling – documents, wall charts, computer-based CASE-tools, and so on; physical size and amount and effort to manipulate them.
 * Empirical: variety of elements distinguished; error frequencies when being written and read by different users; coding (shapes of boxes); ergonomics of computer–human interaction (CHI) for documentation and CASE tools.
 * Syntactic: languages, natural, constrained, or formal, logical and mathematical methods for modeling.
 * Semantic: interpretation of the elements of the model in terms of the real world; ontological assumptions; operations for arriving at values of elements; justification of external validity.
 * Pragmatic: roles played by models – hypothesis, directive, description, expectation; responsibility for making and using the model;conversations needed to develop and use the model.
 * Social: communities of users; the norms governing use for different purposes; organizational framework for using the model.

These layers can be divided into two groups to indicate the technical and the social aspects. Physical, Empirical and Syntactics forms an area where technical and formal methods are adequate. However Semantics, Pragmatics and the Social sphere cannot be explained using those methods unchanged. This points to the fact that one has to apply human judgement when discussing quality in the higher semiotic layers. To solve this issue a generic quality framework, SEQUAL (Krogstie et al 2006) has been developed. The main concepts of the SEQUAL framework and their relationships are illustrated in Figure 3.5.

Fig 3.5: SEQUAL: Framework for discussing the quality of models

The language quality relates the modeling language used to the other sets. Six quality areas for language quality are identified with aspects related to both the language metamodel and the notation as illustrated in figure 3.6

Fig 3.6: Language Quality in SEQUAL

Process Design and Engineering
Work processes models have long been utilized to learn about, guide and support practice. In software process improvement (Bandinelli et al.1995; Derniame 1998), enterprise modeling (Fox and Gruninger 2000) , and quality management, process models describe methods and standard working procedures. Simulation and quantitative analyses are also performed to improve efficiency (Abdel-Hamid and Madnick 1989; Kuntz et al. 1998). In process centric software engineering environments (Ambriola et al. 1997; Cugola 1998) and workflow systems (WfMC 2000), model execution is automated. This wide range of applications is reflected in current notations, which emphasize different aspects of work. Carlsen (1998) identifies five categories of process modeling languages (PMLs):transformational, conversational, role-oriented, constraint-based, and systemic. The increased interest in modeling processes with UML(Marshall 1999) requires that object-oriented process modeling also be discussed.

Organizational development and learning
According to Argyris and Schön (1996), the prevalent models of reality in the organization influences learning. The prevailing models influences how the organization acts upon itself and its environment in general. Models of reality exist as organizational knowledge that is embedded in routines and practices which may be inspected and decoded even when the individuals who carry them out are unable to put them into words.The authors denote such knowledge as theories of action, which may have two different forms, the contents of which do not necessarily coincide. Espoused theory is the “official” theory used in explaining or justifying a way of performing some activity, whereas theory-in-use is the “real” theory governing activity,found implicit in the performance. The latter may be partly tacit and thus needs to be constructed from observation (Argyris and Schön 1996). Organizational learning implies organizational inquiry (which implies uncovering the theories that govern activity)into the models that govern practice within the organization.Argyris and Schön identify single-loop and double-loop learning as two conceptually different processes,distinguished by the extent to which existing models are changed as part of learning. Stated briefly, single-loop learning results in the learner solving the same problem in another way; a more effective or efficient one if the learning is productive. Double-loop learning implies that the problem is reframed in accordance with higher-order goals or strategy. The result of the learning process is the ability to solve new problems differently, preferably in a way serving the overall strategy in a more effective or efficient way than before.

Product Design and Engineering
According to Lillehagen and Krogstie (2008), most design and modeling projects start from quite abstract and general views, far out from the core knowledge of the targeted results. Core knowledge according to the authors is the metaknowledge supporting and created by performing work.Modeling is about capturing, structuring, and evolving this knowledge. Industrial knowledge to be collected and captured is influenced by legacy thinking and sources such as paper, drawings, and natural language.Industry is still struggling to support product and process life cycle design. The Product Organization Process and Systems (POPS) methodologies will support a holistic design approach. Parnas’ theories about the industrial challenges of classes and decomposing properties into parameters, parameters into attributes, and managing attributes with multiple value-sets have inspired the development of a holistic approach to enterprise design and development. The holistic approach provided by AKM will enable much of what Parnas described, supporting sets of parameters for desired, max./min., actual and last agreed values, and value aggregation and propagation.

Systems Engineering
According to Lillehagen and Krogstie (2008) ,In modeling there has been a long-time trend on supporting the development of new modeling languages (so-called metamodeling) rather than the use of existing languages.Generally accepted conceptual framework for metamodeling explains the relationships between meta-meta-model, meta-model, model, and (not completely correctly named) “user data.” Together they form four layers on top of each other, illustrated in metaobject facility (MOF) in OMG (which again is based on the work on CDIF in the 1980s):
 * The user object layer is composed of the information that we wish to describe. This information is in a database-world typically referred to as “data,” but this is just as much a model as any of the other levels. More precisely, it is a model on the instance level.
 * The model layer is composed of the metadata that describes information. Metadata are informally aggregated as models.
 * The metamodel layer is composed of the descriptions (i.e., meta-meta-data) that define the structure and semantics of meta-data. Meta-meta-data is informally aggregated as meta-models. A meta-model can also be thought of as a “language” for describing different kinds of data.
 * The meta-meta-model layer is composed of the description of the structure and semantics of meta-meta-data. In other words, it is the “language” for defining different kinds of meta-data.

In EXTERNAL (2003), another four level model was proposed. The four levels identified in EXTERNAL are:
 * Layer 1 – Describe process logic: At this layer, one identifies the constituent activities of generic, repetitive processes, and the logical dependencies between these activities. A process model at this layer should be transferable across time and space to a mixture of execution environments.
 * Layer 2 – Engineer activities: Here process models are expanded and elaborated to facilitate business solutions. Elaboration includes concretization, decomposition, and specialization. Integration with local execution environment is achieved.
 * Layer 3 – Manage work: The more abstract layers of process logic and of activity description provide constraints, but also useful resources (in the form of process templates) to the planning and performance of each process. At layer 3, more detailed decisions are taken regarding the performance of work in the actual work environment with its organizational, information, and tool resources; the scope is narrowed down to an actual process instance. Concrete resources increasingly are intertwined in the model, leading to the introduction of more dependencies. Management of activities may be said to consist of detailed planning, coordination and preparation for resource allocation.
 * Layer 4 – Perform work: This lowest layer of the model covers the actual execution of tasks according to the determined granularity of work breakdown, which in practice is coupled to issues of empowerment and decentralization. When a group or person performs the task, whether to supply a further decomposition may be left to their discretion, or alternative candidate decompositions might be provided as advisory resources. At this layer resources are utilized or consumed, in an exclusive or shared manner.

Principles for AKM
According to Lillehagen and Krogstie (2008), AKM regards business knowledge to be the main innovative and integrating force. In order for IT to facilitate harvesting and cultivation of business knowledge, it must be driven by pragmatic representations of people's knowledge. The only way to achieve this is to make end users define, manage and own their active knowledge models. This requires a new way of representing knowledge as visual structures with simple and agile business concepts. To leverage this potential the authors introduce 31 core principles for AKM. These principles are all build into the Enterprise Knowledge Architecture (EKA). Some of the principles for AKM are as follows:   A model is a constellation of multiple views  Related views are mutually reflective  Views capture different dimensions of reality (as aspects) <li> Views from different perspectives may seem to be inconsistent <li> Different perspectives will define different model structures and hierarchies (types and parts) <li> Metamodeling is modeling, and all elements are inherently reflective <li> Any model element can have a multitude of types (including basic types such as objects, relationship, property, etc in different views) <li> Explicit classification should be complemented by implicit and derived classes <li> Property is a fundamental modeling construct <li> Properties anchor evolving parameter trees and value sets <li> Relationship is a fundamental modeling construct <li> Relationships represent complex task patterns <li> Value is a fundamental modeling construct. Values can be related to other elements, having properties, etc </ol>

Core Components of an AKM-built Platform
The core components of an AKM-built Platform are as follow:
 * Its core modeling languages – POPS and IRTV
 * Its approach (C3S3P – concept testing, scaffolding, scenario modeling, solutions modeling, platform configuration, platform delivery and practicing, performance improvement and operations)
 * Its methodologies (CPPD)
 * Its enterprise knowledge spaces and the generic EKA (the enterprise of all enterprises)
 * Its customer-specific AKA
 * Its reconfigurable user-composable services – Model-configured,User-composed Platform Services (MUPS)

AKM Approach
According to Lillehagen and Krogstie (2008), The AKM approach has at its core a customer delivery process with seven distinct steps. The AKM approach is also about mutual learning, discovering, externalizing, and sharing new knowledge with partners and colleagues; knowledge that neither you nor they knew they possesed. Tacit knowledge is most vividly externalized by letting people who contribute to the same end product actually work together, all the time exchanging, capturing, and synthesizing their views, methods, properties, parameter trees and values, and validating their solutions. The seven distinct steps are defined as follows:
 * Concept Testing - it is about creating a customer's interest and motivation for applying the AKM technology. This is done by running pilots and assessing value propositions and benefits from applying the AKM approach
 * Scaffolding is purely about expressing shareholder information structures and views, and relating them to roles, activities and systems to provide a model to raise the customer's understanding for modeling and inspire motivation and belief in the benefits and values of the AKM approach.
 * Scenario Modeling is about modeling the best - practice work processes. Capturing the steps and routines that are or should be adhered to when performing the work they describe. This is the core competence of the enterprise, and capturing these work - processes is vital to perform work, support execution, and perform several kinds of analyses in the solutions modeling step.
 * Solutions Modeling is about cross-disciplinary and cross-functional teams working together to proactively learn and improve quality in most enterprise life cycle aspects. The purpose is creating a coherent and consistent holistic model or rather structures of models and submodels meeting a well-articulated purpose. Solutions modeling involves top-down, bottom-up, and middle-out multidimensional modeling for reflective behaviour and execution.
 * Platform Configuration is about integrating other systems and tools by modeling other systems data models and other aspects often found as UML models. These are created as integral submodels of the customized AKM platforms, and their functionality will complement the Collaborative Product and Process Design (CPPD) methodology with Product Lifecycle Management(PLM)system functions, linking the required web-services with available software components.
 * Platform Delivery and Practicing adapts services to continuous growth and change by providing services to keep consistency and compliance across platforms and networks as the user community and project networking expands, involving dynamic deployment of model-designed and configured workplace solutions and services.
 * Performance improvement and operations is continuously performing adaptations or providing services to semi-automatically reiterate structures and solution models, adjusting platform models and regenerating model-configured and generated workplaces and services, and tuning solutions to produce the desired effects and results.

CPPD Methodology
The CPPD methodology provides methods, languages and workplaces for enterprise and product designers and other stakeholders to effectively and continuosly perform collaborative business, design and engineering. CPPD methodology provides languages and methods for improved stakeholder involvement, design, engineering and business interactions and knowledged and data sharing from the onset of any new innovation or delivery project.

CPPD Components
According to Lillehagen and Krogstie (2008), CPPD components are delivered to customer projects as a set of generic knowledge models and services, allowing and guiding cross-partner service-teams in developing generic and adapting project specific enterprise knowledge models, methods, and rules, thereby creating application platforms and services to meet customer specific demands. There are currently 12 configurable CPPD components defined. These components are developed, applied, and operated as a coherent, consistent, and compliant set of generic services and reference architectures for capturing, representing, and reusing enterprise knowledge to support holistic product design. The 12 configurable CDDP componennts currently defined are listed below: <ol> <li>Configurable Product Components (CPC) is used for capturing parameterized variants, shapes, and materials. <li>Configurable Visual Workspaces (CVW) is used for designing and generating user workspaces. <li>Configurable Work Processes (CWP) is used for managing dependencies between task. <li>Configurable Property and Parameter-Sets (CPP) makes it possible to handle properties, and parameters separately by each engineering or business discipline <li>Configurable Product Structure (CPS) is an early design support language for generic model and services <li>Configurable Function Deployment (CFD) is used to correlate requirements and constraints with product properties and features <li>Configurable Design Language (CDL) is used for linking conceptual EKA to sketches illustrating fundamental and innovative product concepts <li>Configurable Idea Bank (CIB) is used for capturing and relating design ideas, principles, requirements, sketches, constraints, and stakeholder views for more effective innovation <li>Configurable Web service Integration (CWI) used for interfacing legacy systems as web services <li>Configurable Web Workplaces (CWW) is used for designing and generating workplaces on the web <li>Configurable Collaboration Spaces (CCS) used for configuring roles, tasks and views <li>Configurable Competence and skill Profiles (CCP) used for visual competency management </ol>

Main Industrial Challenges
Individuals and organizations have been faced with a lot of challenges such as large amount of information that is not yet digitally available, lack of support for shared business models, business interoperability and methodologies for inter-enterprise collaboration.Listed below are the major industrial challenges <ol> <li>Building searchable digital information libraries of present common information sources, to improve data and knowledge sharing and use. <li>Developing consistent reference models that are easily integrated with Web-platforms, to allow more effective community and project extensions and adaptations. <li>Developing knowledge engineering platforms and services that can add value to and integrate present IT application systems, “the islands of automation” <li>Developing operational enterprise knowledge architectures and platforms to concretize and make operational current blueprint architectural frameworks <li>Develop methodologies as descriptive templates to support the building of industry platforms, for example the CPD methods to build collaborative design platforms <li>To model reference models that can be reused and drive knowledge standardization initiatives across projects and sectors <li>To support holistic design implying that multidimensional modeling capabilities to express mental models of designers and engineers must be supported <li>To provide modeling team services and role-specific workplaces and views to support concurrent knowledge engineering for collaborative product design <li>To provide model or knowledge architecture configured workplaces to enable new approaches to model-based systems engineering and solutions deployment <li>To provide services to enable data definition and sharing without being dependent on IT-defined data-models, thus supporting idea capture and conceptual design. </ol>

Impact of AKM Technology Relative To Challenges
By implementing and deploying AKM technology, industries, individuals, Sciences and other stakeholders have the benefit of exploiting knowledge technology and IT for improved human creativity and more pragmatic learning. According to Lillehagen and Krogstie (2008) the AKM technology will enable industrial users to build their own operational networks, workplaces as well as collaboration arenas, this will enable global collaboration and improve most industrial practices and sciences. An Active Knowledge Architecture covering the early design stages has the capabilities of cutting down on both cost and time by factors. Another general impact from the AKM technology would be convergence of scientific concepts and disciplines.The AKM technology has the capability of describing any role in an industry making use of the visual modeling using the IRTV language and the C3S3P methodology. The AKM technology will aid industries to be able to effectively handle resource-consuming change and version management.

AKM technology provides visual language definition services to enable industrial designers to dynamically define evolving product artifacts, combining object instances, properties, and task patterns by capturing the designer actions as an integrating task pattern.