قسم االنشاءات كلية الهندسة جامعة المنصورة Knowledge Management Cairo University 6 th December 2016 أ.د./ ابراهيم مطاوع E-mail: i_a_motawa@mans.edu.eg
Learning Outcomes Recognise the meaning, nature and importance of knowledge management in construction organisations Examine the role of management information systems (MIS) in the management of knowledge
What is Knowledge? Term Data Information Knowledge Description Values obtained from any type of sensors, including humans. Geographical analogy: heights, distances, etc. Synthesised from data, to produce a map of the territory (information is the potential for knowledge) What the map means, how it can be used for different purposes (importance of purpose ), when it should not be used, when it needs change, if it is insufficient, or when it needs link with other information.
Properties of Knowledge Includes information and data, but these, on their own, may not include knowledge Should have purpose (e.g. adding value through its application) It is both time and context sensitive Formal aspects of this knowledge (not tacit aspects) can be taught Tacit aspects tend to reside in human heads and can be difficult to capture or formalise It can be costly to generate/replace, either tacit or explicit
Types of Knowledge Explicit (Formal) & Tacit Knowledge Knowledge for business relevance and functional role within an organisation
Explicit (Formal) Knowledge Easier to identify Reusable in a consistent and repeatable manner May be stored as: written procedure, process in computer system, etc.
Tacit Knowledge (Expertise) What is held in human heads (understanding, implied, existing without being stated) Difficult to transfer The interface between formal knowledge and its application to real problems knowledge that oils the wheels of formal procedures
Knowledge conversion
Knowledge for Business Relevance and Functional Roles With respect to the role of knowledge within an organisation: Knowledge of people (the behaviour of stakeholders/clients/suppliers, relationships and purposes) Business environment insights Knowledge embedded in processes, products, services, etc. (Organisation memory)
Why manage knowledge? The importance of knowledge to organisational success Emergence of information economy Importance for competitive advantage knowledge & core competencies (the few things an organisation does best) are key organisational assets
What is Knowledge Management? Aim of KM is the recognition of the strategic value of its intellectual assets and the careful management and distribution of these assets across the enterprise to create value, increase productivity and gain and sustain competitive advantage A set of processes to capture, preserve and disseminate the knowledge of key individuals or groups in the organisation to assure the availability of that knowledge later when the individual has retired or the groups have disperse KM involves: Generation (creation of new knowledge) Capture of existing knowledge Storage (humans, database, tools) Accessibility (Registry and search mechanisms) Application Dissemination Retirement of knowledge In general KM seeks for: Explicit knowledge: consolidation and making available of artefacts Tacit knowledge: creation of communities to hold, share and grow tacit knowledge
IT infrastructure for KM Create Knowledge (Knowledge Work Systems) CAD Virtual Reality Distribute Knowledge (Office Automation Systems) Desktop Publishing Imaging Electronic calendars Desktop Databases Share Knowledge (Group Collaboration Systems) Groupware Intranets Capture & Codify Knowledge (Artificial Intelligence Systems) Expert Systems Neural Nets Fuzzy Logic Genetic Algorithms Intelligent Agents
Knowledge Work Systems (KWS) An information system that aids knowledge workers in the creation and integration of new knowledge in the organisation They must give knowledge workers the specialised tools they need (e.g. powerful graphics, analytic tools, communications and document management tools great computing power with quick and easy access to external databases) Computer-aided design (CAD) automate the creation and revisions of designs using computers and sophisticated graphics software Virtual Reality Systems have visualisation, rendering and simulation capabilities that surpass CAD systems use of interactive graphics to create computer-generated simulations
Knowledge Distribution Connecting the work of the local knowledge workers with all levels and functions of the whole organisation and the external world, including customers, suppliers, government regulators, and external auditors Management of documents (creation, storage, retrieval and dissemination) Communicating including initiating, receiving, and managing voice. Digital and document-based communication for both individuals and groups. Office Automation Systems (OAS) Any application of information technology that intends to increase productivity of knowledge workers in the office e.g. word processing, voice mail, video conferencing digital image processing (imaging systems) at the core of current OAS group assistance tools (e.g. networked digital calendars) to assist group work among office workers
Knowledge Sharing Necessary to facilitate group work: coordination and collaboration Tools include: email, teleconferencing, data conferencing, groupware, Internet, etc. Note: group collaboration technologies alone cannot promote sharing if team members do not want to share knowledge
Knowledge Capture and Storage Use of artificial intelligence (AI) to: capture individual and collective knowledge codify and extend organisational knowledge base AI systems: are efforts to develop computer-based systems that behave as humans preserve expertise that might be lost through the absence of an expert store knowledge in a active form that can be accessed by others create mechanism not subjected to human feelings eliminate routine and unsatisfying jobs held by people Successful AI systems are based on human expertise, knowledge, and selected reasoning patterns but they do not exhibit the intelligence of human beings They do not invent new and novel solutions to problems AI applications include: Robotics, Expert systems, intelligent machines, etc.
Expert Systems Knowledge intensive systems that capture human expertise in limited domains of knowledge Assist in decision making by asking relevant questions and providing explanation for action taken Can help organisations make higher level decisions with fewer people Quite narrow, shallow and brittle
How Expert Systems work Human Knowledge Knowledge Frames Rules Income Salary Consultancy Savings interest Expenses Housing Tax Holidays Budget Income Expenses If-Then-Else If condition is true, then an action is taken, else another action is taken
Simple example to select the best contractor A If turnover > 500,000 Ask about Years in Business Else EXIT B If Years in Business > 10 years, Ask about No. of employees Else EXIT C If No. of employees is between 75 and 100, Grant one project Else EXIT D For on going projects Ask about performance E If missing project deadline > 5% of project duration, Do F (grant another one project) Else Do G G If missing project budget < 3% Ask about other debt Else EXIT H If other debt >5% of turnover, Do F Else Do I F Only one more project I Agree a long term partnership
Development of Expert Systems Developed within a shell (a programming environment) e.g. Kappa PC, Programming Language Prolog, etc The main contributors: Expert: have thorough command over knowledge base Knowledge Engineer: Special expert in eliciting information/knowledge from other professionals Translates information/knowledge into set of rules and/or frames for an expert system
Development of Expert Systems The process involves: Selecting/ensuring that the problem is appropriate for an expert system Determine balance between potential savings from system against the cost of developing it Develop prototype system to test assumptions about how to encode the knowledge of experts Develop a full-scale system, with specific focus on the addition of a very large number of rules Check the comprehensibility of system and prune, if necessary to achieve simplicity and power Test system by a range of actual experts within the organisation against any performance established earlier (e.g. output of the system should agree with that of experts for 90% of the time, etc.) After successful testing, integrate system into data flow and work patterns of the organisation
Problems with Expert Systems Lacks the robust and general intelligence of humans Suitable for only certain classes of problems and represent limited forms of knowledge Development efforts can be very long Knowledge base is fragile and brittle - cannot learn to change over time Based on prior knowledge of a few known alternatives
Case-Based Reasoning Unlike expert systems which work by applying a set of IF-THEN-ELSE rules, CBR represents knowledge as a series of cases and this knowledge base is continuously expanded and refined by users Represents knowledge as a database of cases for later retrieval when a similar case is encountered System searches for stored cases similar to the new one, finds the closest fit, and applies the solutions of the old case to the new case Successful solutions are tagged to the new case and both are stored together Unsuccessful solutions are also appended to case database with explanations on why they did not work
How Case-Based Reasoning works 1 User describes the problem 2 System searches database for similar cases 3 System asks user additional questions to narrow the search Case database 4 System finds closest fit and retrieves solution 5 System modifies the solution to better fit the problem System stores problem and successful solution in the database 6 NO Successful? YES
Artificial Neural Networks (ANN) Attempt to emulate the processing patterns of the brain - learning by example Consist of an input, output and a hidden processing layer It learns from the input/output patterns of data to construct a hidden layer of logic as follows: network is fed training data for which inputs produce known outputs. This helps the computer to learn the correct solution by example as more data is fed, each case is compared with known outcome: if it differs, correction is calculated and applied to nodes in hidden processing layer Steps are repeated until a satisfactory condition, such as corrections being less than a certain amount, is reached
Neural Network structure an example Neural Network Input Layer Hidden Layer Output Layer Turnover Debt Years in Business Grant one project Agree partnership Performance
Problems of ANN Cannot always explain their outputs Cannot guarantee certainty Is sensitive to the training data (Too little or too much data)
Class Activity 1 Given a work programme of a construction project that shows when each task starts and finishes. It also shows the resources required. There are potential interruptions to the programme (e.g. delay, resource shortage, unforeseen events, etc.). Give an example of solution that a project manager can consider. Your example should show relevance to the tacit knowledge of this manager.
Class Exercise 2 Consider the problem of programme interruption for the construction process, use expert systems rules to develop IF- THEN-ELSE statements for its solution