Multi attribute decision making


Published on

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Multi attribute decision making

  1. 1. Name: Shriraam Madanagopal<br />Distance Student<br />TERM PAPER<br />Engineering Economy –IE 5304<br />Topic: Multi Attribute Decision Making<br />Multi Attribute Decision Making<br />Introduction:<br />Multi-attribute decision making method, has its base for the decision making model, is one of the decision-making support methods. The decision making model is based on a chosen list of criteria, parameters, variables or factors, which we wish to monitor in the decision making process. The formal base for the establishment of a model-Multi attribute decision making, in which the key criteria is the interconnectedness of assessments according to the individual parameters that result in an integrated assessment. From the observations of the case study, a better understanding of the advancement in Multi Attribute Decision Making model and cognitive techniques underlining the observed effects and precise information with respect to the analysis of a project uses task variants. The Multi Attribute Decision support can be best explained as a procedure aimed at the evaluation of options that occur in decision making process. This procedure will take us through a description of the scenario and a study of the results.<br />Decision Models:<br />One of the basic approaches to the decision support is the multi attribute decision making with the basic principle of decomposing the decision problem into smaller, less complex sub-problems. Options are decomposed onto various dimensions X, usually called the attributes, parameters or criteria. After decomposing, each option O is defined by a vector value v for the corresponding attributes. These vectors are evaluated by a Utility Function F. The Utility Function should be well defined by the decision maker(s), by defining his or her or their main objectives or goals. When these rules or protocol is applied to the particular option O, the function F yields to a Utility F(O). According to this value, the options can be ranked as per the decision maker’s goals and the best one is selected. The attributes X and the Utility Function F are the two main factors which determines the decision maker’s knowledge about a particular decision problem. Added to this there is a data base of options, consisting of vectors v. The procedure of hierarchical models has been developed and extensively applied in relation to the decision support. <br />The attributes are represented in the form of a tree which gives the outline for the decision problem. These attributes are designed according to their interdependence. The leaves of the tree are referred as basic attributes, which solely depend on the characteristic of options. The internal nodes of the tree are called the aggregate attributes and their values are determined by the basis of the Utility Functions. The root of the tree is the most important aggregate attribute. These roots represent the overall Utility of options.<br />These Utility functions define the process of aggregation of lower level attributes into the corresponding higher –level fathers. Every aggregate attribute X, a Utility function F that maps the values of sons of X into values of X, must be defined by the person who makes decision. <br />Utility functions are represented by elementary decision rules.<br />Let X1, X2, X3…..Xk be the predecessor of an aggregate attribute Y. Then the Function Y= F( X1, X2,X3….. Xk) be defined by a set of rules of the form:<br />X1= x1 and Xk = xk and<br />Y = ym : yM, where xi, ym and yM represent the values of the Corresponding attributes. “ym : yM” stands for an interval of values between ym and yM, inclusive.<br />The most commonly, ym:yM is a single-value interval. In this case, the rule is simplified to<br />if X1 = x1 ….. Xk = xk then Y = y. <br />Tables are formed by grouping set of elementary decision rules. When more decision making groups with different objectives are involved on the decision process, each group can define their own set of utility functions.<br />Figure a: General Concept of multi attribute Decision making<br />Utility<br /> F(O1)…F(Ok)<br /> F<br />Utility Functions<br /> Attributes<br />X1 X2 …Xk <br /> O1= ( V11 V12 … V1n) Options<br /> Ok= ( Vk1 Vk2 .... Vkn)<br /> <br /> Multi Attribute Decision Model: <br />Multi Attribute Decision Model can be developed with qualitative hierarchical decision model which is based on the DEX shell. This helps to create the decision models that consist of non-numerical qualitative criteria. The criteria are hierarchically ordered into a tree structure. The weights are replaced by rules that define the interdependence of the criteria and their influence on the final evaluation. Thus the influence of criteria can depend on its value, which corresponds in Utility theory to the variability of weights. The qualitative hierarchical decision model is based on a chosen list of criteria, parameters, variables or factors, which we are going to monitor in the decision making process.<br />The Decision making process is divided into five phases:<br /><ul><li>Identification of the problem
  2. 2. Criteria identification and criteria structuring
  3. 3. Utility function(decision rules)
  4. 4. Description of variants
  5. 5. LMS Evaluation and analysis</li></ul>Incorporation of Multi Attribute Decision making in design and manufacturing project:<br />The analysis of this Circlip project started with my design interest since my under graduation days in India pistons limited. I come from a Mechanical Engineering background in which I wanted to study about the various design factors and optimizing the various constraints involved while designing an industrial component. At that point of time I wasn’t introduced to the Engineering Economy and knowledge engineering concepts. But now with the Engineering Economy and Knowledge Engineering background which made me work on the concept of designing an optimal Circlip by taking into consideration the constraints involved like material costs, manufacturing costs and type of material used and various dimension oriented concepts. I would consider this subject as a boon which helped me in making my designing dreams come true where Optimization plays a very vital role. Data analysis and processing are very important and vital in every aspect of life. Be it a normal office work or a big research industry we should be in a position to know how and why each and every decision is made. This kind of a learning process makes an engineer or a doctor or any professional to get a grasp on what he or she is doing.<br />The Circlip:<br />The circlip is a locking mechanism which is fastened on to the gudgeon pin. It is a very integral and important part of the Piston assembly. The circlip seats in a groove known as the circlip groove. The circlip groove is the one that restrains the motion of the gudgeon pin. If there wasn’t a concept of circlip, then the gudgeon pin would move freely, colliding against the wall increasing friction and engine damage along with causing damage to itself. In order to prevent this damage, the circlip is used. Though a very small component in the piston assembly, it plays a vital and integral role in the performance and the life of the piston assembly. They have to be machined very accurately. The groove though a small groove, it is to be machined with great tolerance limit.<br />The Circlip Grooving: <br />The Circlip groove is the groove in which the circlip sits. The groove has to be manufactured and cut with precision because even small changes in the groove dimension can cause movement of the circlip within the groove. This movement will cause the gudgeon pin to vibrate causing wear and tear on the piston walls and on the gudgeon pin. The Circlip grooving operation is an operation where the groove, which is being cut, is an undercut, which means that the tool has to move into the piston and perform the grooving operation.  <br />The piston, which is being manufactured, is made of an Aluminum alloy, which is molded with Cast Iron (CI).  The existing Circlip Grooving process is carried out on a Pillar drilling machine. The pillar-drilling machine is a drilling machine, which has the capacity to drill larger material pieces and can drill larger holes as well. The tool is fixed on to the Drilling machine with the help of a NOBUR head. The NOBUR head contains the mechanism that assists the tool movement during rotation to perform the grooving operation.  <br />Identification of the Problem:<br />With respect to the fact that the human resource deployment has been reorganized as one of the most important elements for further development of modern societies, the current demand for new knowledge skills has constantly increased. Learning Management Systems (LMSs) which support e-learning have been developed and are available on the market. Learning Management Systems (LMS) are systems that support the creation (via authoring tools), storage (for example in a relational database) and presentation (often via a web browser) of learning materials in a structured way. They also include ‘tracking’ tools that allow for record-keeping on students enrolled in courses, and usage statistics for the system as a whole (one of the most important of these being statistical analysis of students’ responses to questions, which enables validation of<br />testing on the system).<br />Designing the Circlip with various problems like:<br />Stress Conditions(what stress can the circlip take subject to worst possible conditions)<br />Allowable stress = The Ratio of the Yield Strength to the Factor of safety.<br />Yield Strength is the stress at which the material begins to deform.<br />Factor of Safety, the maximum expected load to which a component or an assembly is subjected to.<br />Materials used in making the circlip .In order to make sure that the brittle factor is taken into consideration, we can improve the efficiency of the overall performance of the piston.<br />Material cost = Total weight of material used multiplied by the cost per kg of material.<br />Manufacturing Cost: Forging, Bending and Cutting costs respectively.<br />Values of stress, strain and cost for each parameter<br />Identification, Description and Criteria Structuring:<br />This level explains the descriptions of criteria which are the components of the decision making model. When we create a model we must take into consideration to meet the requirements of the design project. In order to meet the requirements of this project, we introduce the principle of criteria integrity (Inclusion of all relevant criteria), appropriate structure, comprehensiveness, non-redundancy and measurability. <br />Comprehensiveness –checks for all relevant data is present in the data base.<br />Non-redundancy means each individual piece of data exist only once in the data base.<br />Appropriate structure means that the data are stored in such a way as to minimize the cost of expected processing and storage.<br /> <br />The criteria are divided into three scopes:<br />The design procedure, cost aspects and the material science of the whole project forms the skeleton of the multi attribute model. The criteria can include the following values: ‘Low’, ‘Average’ or ‘High’. The only exception being the criteria where it is impossible to determine an intermediate value. All values have an increasing range –low value is the worst than the higher value. The first group Criteria is merged into the design procedure category composed of four basic attributes: ease of design, design factors, designing tools and designing materials. These aspects are taken into consideration and we make sure that all the design aspects communicate with each other and no chaos establishment is tried using this stage. Similarly the other two criteria are taken care of.<br />Utility Function:<br />The tree of criteria defines the structure of the evaluation model by defining the criteria and the interdependence. In the final outcome, this means that the overall evaluation of LMS’s depends on Optimization criteria of decision making. On the other hand, the criteria tree does not define the aggregation- the procedure is defined by decision rules.<br />The rules determine the evaluation of the criterion, ease of design, design factors, designing tools and designing materials. The first five rules determine the conditions by which the design procedure is evaluated as unsuitable (low grade). This is for example whenever the LMS does not conform to ease of design, design factors, designing tools (regardless of the evaluation of the remaining criteria. On the other hand if the design procedure is suitable-high value, whenever LMS respects the ease of design at least on an average level-average grade and the quality of the attributes design factors, designing tools and designing materials are high. The remaining rules can be interpreted similarly, with the symbols <= and >= representing " worse or equal" and " better or equal" , respectively.<br />S.No3829211291LowLow <= Average *Low 2LowLow*LowLow 3LowLow LowLowLow 4Low<= AverageLow *Low 5Low <= Average<= Average *Low 6Average <= AverageLowLowLow 7Average High High HighHigh 8High >= Average>= AverageHighHigh 9High >= Average **High 10High High High HighHigh 11High High>= Average *High 12High HighHigh HighHigh<br /> Ease of design, design factors, designing tools and designing materials<br /> <br />The table above represents the Utility criteria function for design procedures<br />Obviously there are many more rules in the model. For each aggregate criterion, a similar table can be designed. Experts contribute to the contents of the rules, by covering all possible combinations of the evaluation criteria and consistent improvement single lower level criterion should never decrease the overall value of the LMS. Decision rules therefore define the conditions under which an LMS is ranked.<br />Description of variants:<br />The multi-attribute decision making model was tested on three LMSs: Blackboard 6 (,<br />CLIX 5.0 ( and Moodle 1.5.2( Blackboard is among the most<br />perfected and complex LMSs on the market. The system offers various communication options (both synchronous<br />and asynchronous) within the learning project. The Blackboard LMS is designed for institutions dedicated to teaching and learning. Blackboard technology and resources power the online, web-enhanced, and hybrid education programs at more than 2000 academic institutions (research university, community college, high school, virtual MBA programs etc. CLIX is targeted most of all at big corporations, because it provides efficient, manageable, connected and expandable internet-based learning solutions. This scalable, multilingual and customizable software aims at providing process excellence for educational institutions. For educational administrators, CLIX offers powerful features for course management and distribution. <br />LMS Evaluation and Analysis:<br />Once the knowledge base has been built, the second main part of DEX, i.e evaluation and analysis of our project options, can be applied. At the beginning, the user activates a specialized editor of options in order to describe the options by assigning the corresponding values to the basic attributes. After this the DEX automatically evaluates our options as follows:<br />Interactive inspection: by walking around the tree and looking at the values that were assigned to aggregate attribute during the evaluation.<br />Explanation of the evaluation: DEX can explain how each particular value has been obtained in terms of attributes' values involved in the process, triggered rules and descriptions of computations performed by DEX.<br /> What-if analysis: is performed interactively by changing values of basic attributes, reevaluating options and comparing the obtained results with the original ones.<br />Selective explanation of options: DEX finds and reports those sub trees that expressed the most<br />advantageous or disadvantageous characteristics of a particular option. The main point is in the explanation of options using only the most relevant information. In the design of DEX, one of the most important goals was the transparency to the user. For this reason, the user can access a powerful report generator during all the stages of working with DEX. The generator is able to prepare complete or partial reports showing the components of the knowledge base, evaluation/analysis results and different kinds of explanation. The user can choose between different levels of detail and different forms of representation. The reports can be inspected interactively or printed out.<br />The evaluation of LMSs is carried out according to the tree of criteria from the basic criteria up. The method of aggregation is determined by the decision rules. The variant which is awarded the highest grade should be the best one.<br />Result:<br />With the help of multi attribute decision making we can make an optimal, cost efficient design and fabricate a universal tool for Circlip Grooving Operation by taking into consideration various parameters as discussed above. Both theory and practice are becoming more focused on human-computer interactions. They are focusing more on the applicability, usability and adequacy of Learning Management Systems. Their theory and practice base their assumptions on the project’s environment. Thus our project will be successfully established with the help of multi attribute decision making model.<br />References:<br />R. Johnson, J. R. Hegarty, “Websites as Educational Motivators for Adults with Learning Disability”, British<br />Journal of Educational Technology, vol. 34(4), 2003, pp.<br />479–486.<br /> <br />M. Bohanec, V. Rajkovič, ”Multi-Attribute Decision Modeling: Industrial Applications of DEX”, Informatica,<br />vol. 23, 1999, pp. 487–491.<br />V. Chankong, Y. Y. Haimes, Multi objective Decision Making: Theory and Methodology, North-Holland, 1983<br />P. G. W. Keen, M. S. Scott Morton, Decision Support<br />Systems – An Organizational Perspective. Addison-<br />Wesley, 1978.<br /> S. L. Alter, Decision Support Systems. Current Practice and<br />Continuing Challenges, Addison-Wesley, 1980.<br />E. Turban, Decision Support and Expert Systems. New<br />York: Macmillian, 1988.<br />S. French, Decision theory: An Introduction to the<br />mathematics of rationality. New York: Wiley, 1986<br />J. N. Mandić, H. E. Mamdani, “A multi-attribute decision making model with fuzzy rule-based modification of priorities”, in Zimmerman, Zadeh, and Gaines (Eds.),<br />Fuzzy Sets and Decision Analysis, Elsevier Publishers,<br />North-Holland, 1984, pp. 285-306.<br />K. Keenoy, G. Papamarkos, Learning Management Systems and Learning Object Repositories. Birkbeck College, University of London, 2003.<br />L. D. Phillips, Decision analysis and its applications in industry. In G. Mitra, editor, Computer Assisted Decision Making, Elsevier (North-Holand), 1986, pp. 189–197.<br />R. L. Keeney & H. Raiffa, Decisions with Multiple Objectives. John Wiley & Sons, New York, 1976.<br />Fishburn, P.C. (1970): Utility Theory for Decision Making. New York: John Wiley &<br />Sons.<br />