INTRODUCTION TO MANAGEMENT SCIENCE, 13e Anderson Sweeney Williams Martin © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slides by JOHN LOUCKS St. Edward’s University Slide 1 Chapter 14 Multicriteria Decisions Goal Programming Goal Programming: Formulation and Graphical Solution Scoring Models Analytic Hierarchy Process (AHP) Establishing Priorities Using AHP Using AHP to Develop an Overall Priority Ranking © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 2 Goal Programming Goal programming may be used to solve linear programs with multiple objectives, with each objective viewed as a "goal". In goal programming, di+ and di- , deviation variables, are the amounts a targeted goal i is overachieved or underachieved, respectively. The goals themselves are added to the constraint set with di+ and di- acting as the surplus and slack variables. © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 3 Goal Programming Problem (p.651) • Multiple goals : maximum total return, minimum total risk. • In LP, Maximize the total return subject to allowable maximum risk Or Minimizing the total risk subject to guaranteed minimum return. © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 4 Goal Programming Constraint and goals. • Maximum available fund $80,000 • Maximum allowable risk 700 • Minimum total return $9,000 Decision variables U = number of shares of U.S. Oil purchased H = number of shares of Hub properties purchased d1 + , d1 – : deviation from the goal 700 d2 + , d2 – : deviation from the goal 9000 Constraints 25 U + 50 H <= 80000 0.5U + 0.25H = 700 + d1+ – d1– 3 U + 5H = 9000 + d2 + – d2 – © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 5 Goal Programming Objective function Preemptive priorities : P1, P2, . . ., Pk Pi should be should be optimized before Pi+1 is optimized In most goal programming models deviation variables are minimized. Goal programming model © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 6 Goal Programming Graphical solution © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 7 Goal Programming Computer solution If a Goal programming solver is not available, LP solver should be used in a sequence of P1 goal and then, P2 goal, … etc. Higher goal achievement will be inserted in the formulation of lower goal formulation as a constraint. For goal 1 Min. d1+ 25 U + 50 H <= 80000 0.5U + 0.25H – d1+ + d1– = 700 3 U + 5H – d2 + + d2– = 9000 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 8 Goal Programming P1 solution Objective value: 0.000000 Variable Value D1P 0.000000 U 0.000000 H 0.000000 D1M 700.0000 D2P 0.000000 D2M 9000.000 Row 1 2 3 4 Reduced Cost 1.000000 0.000000 0.000000 0.000000 0.000000 0.000000 Slack or Surplus Dual Price 0.000000 -1.000000 80000.00 0.000000 0.000000 0.000000 0.000000 0.000000 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 9 Goal Programming For goal 2 Min. d2– 25 U + 50 H <= 80000 0.5U + 0.25H – d1+ + d1– = 700 3 U + 5H – d2 + + d2– = 9000 d1+ = 0 P2 solution Objective value: Variable Value D2M 600.0000 U 800.0000 H 1200.000 D1P 0.000000 D1M 0.000000 D2P 0.000000 600.0000 Reduced Cost 0.000000 0.000000 0.000000 0.000000 1.333333 1.000000 Row 1 2 3 4 5 Slack or Surplus Dual Price 600.0000 -1.000000 0.000000 0.9333333E-01 0.000000 1.333333 0.000000 -1.000000 0.000000 1.333333 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 10 Scoring Model for Job Selection (p.665) A graduating college student with a double major in Finance and Accounting has received the following three job offers: • financial analyst for an investment firm in Chicago • accountant for a manufacturing firm in Denver • auditor for a CPA firm in Houston © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 11 Scoring Model for Job Selection The student made the following comments: • “The financial analyst position provides the best opportunity for my long-run career advancement.” • “I would prefer living in Denver rather than in Chicago or Houston.” • “I like the management style and philosophy at the Houston CPA firm the best.” Clearly, this is a multicriteria decision. © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 12 Scoring Model for Job Selection Considering only the long-run career advancement criterion: • The financial analyst position in Chicago is the best decision alternative. Considering only the location criterion: • The accountant position in Denver is the best decision alternative. Considering only the style criterion: • The auditor position in Houston is the best alternative. © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 13 Scoring Models Procedure summary • Setup the criteria (kind of goals) • Choose weights of importance and assign the weights to each criterion (subjective weights) (p.666 and Table 14.1 on p.667) • Choose scales to evaluate alternatives for each criterion (p.667) • Evaluate each alternative for each criterion one by one using the chosen scales (Table 14.2 on p668) • Get the weighted average (scores) for each alternative (Table 14.3 on p.668) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 14 Scoring Model: Step 1 List of Criteria (p.666) • Career advancement • Location • Management • Salary • Prestige • Job Security • Enjoyable work © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 15 Scoring Model: Step 2 Five-Point Scale Chosen for criteria (Importance, p.666) Importance Weight Very unimportant 1 Somewhat unimportant 2 Average importance 3 Somewhat important 4 Very important 5 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 16 Scoring Model: Step 2 Assigning a Weight to Each Criterion (p.667) Criterion Career advancement Location Management Salary Prestige Job security Enjoyable work Importance Weight Very important 5 Average importance 3 Somewhat important 4 Average importance 3 Somewhat unimportant 2 Somewhat important 4 Very important 5 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 17 Scoring Model: Step 3 Nine-Point Scale Chosen to evaluate alternatives (p.667) Level of Satisfaction Extremely low Very low Low Slightly low Average Slightly high High Very high Extremely high Rating 1 2 3 4 5 6 7 8 9 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 18 Scoring Model: Step 3 Rate how well each decision alternative satisfies each criterion. (p.668) Criterion Career advancement Location Management Salary Prestige Job security Enjoyable work Decision Alternative Analyst Accountant Auditor Chicago Denver Houston 8 6 4 3 8 7 5 6 9 6 7 5 7 5 4 4 7 6 8 6 5 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 19 Scoring Model: Step 4 Compute the score for each decision alternative. Decision Alternative 1 - Analyst in Chicago Criterion Career advancement Location Management Salary Prestige Job security Enjoyable work Weight (wi ) 5 3 4 3 2 4 5 Rating (ri1) x Score © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. 8 3 5 6 7 4 8 wiri1 = 40 9 20 18 14 16 40 157 Slide 20 Scoring Model: Step 4 Compute the score for each decision alternative. s j wi rij i S1 = 5(8)+3(3)+4(5)+3(6)+2(7)+4(4)+5(8) = 157 S2 = 5(6)+3(8)+4(6)+3(7)+2(5)+4(7)+5(6) = 167 S3 = 5(4)+3(7)+4(9)+3(5)+2(4)+4(6)+5(5) = 149 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 21 Scoring Model: Step 4 Compute the score for each decision alternative. (p.668) Criterion Career advancement Location Management Salary Prestige Job security Enjoyable work Score Decision Alternative Analyst Accountant Auditor Chicago Denver Houston 40 9 20 18 14 16 40 30 24 24 21 10 28 30 20 21 36 15 8 24 25 157 167 149 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 22 Scoring Model: Step 5 Order the decision alternatives from highest score to lowest score. The alternative with the highest score is the recommended alternative. • The accountant position in Denver has the highest score and is the recommended decision alternative. . © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 23 Analytic Hierarchy Process The Analytic Hierarchy Process (AHP), is a procedure designed to quantify managerial judgments of the relative importance of each of several conflicting criteria used in the decision making process. Developed by Saaty(1982), but still used a lot. While scoring model assigns weights to criteria (1~5) and alternatives(1~9) simultaneously considering all the criteria or all the alternatives, AHP assign weights to each pairwise comparison of criteria or to each pairwise comparison of alternatives. © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 24 Analytic Hierarchy Process Problem (p.670) To select a car (Accord, Saturn, Cavalier) based on the characteristics of the cars. Criteria are Price, MPG, Comfort, Style Any criterion can be divided into subcriteria such as Comfort(Interior, Sound system) and Style(Color, Body type) Hierarchy (p.671) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 25 Analytic Hierarchy Process Another Hierarchy (from Wikipedia) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 26 Analytic Hierarchy Process Scale for the Importance of Criteria(p.673) Pairwise comparison of Criteria (p.673) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 27 Analytic Hierarchy Process Pairwise Comparison Matrix for Criteria(p.675) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 28 Analytic Hierarchy Process Normalized Matrix and Matrix with Priority (p.676) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 29 Analytic Hierarchy Process Normalized Matrix and Matrix with Priority (p.676) © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 30 Analytic Hierarchy Process Scale for Preference of alternatives (p.679) Pairwise comparison matrix of alternatives for each criterion © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 31 Analytic Hierarchy Process Priorities for each alternative using each criterion(p.680) Choosing an alternative by calculating weighted average. s j wi rij i Overall Priorities Pr(Accord) = 0.398(0.123)+0.085(0.087)+0.218(0.593)+0.299(0.265) = 0.265 Pr(Saturn) = 0.398(0.320)+0.085(0.274)+0.218(0.341)+0.299(0.656) = 0.421 Pr(Cavalier) = 0.398(0.557)+0.085(0.639)+0.218(0.065)+0.299(0.08) = 0.314 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 32 Analytic Hierarchy Process Procedure • Step 1 : List the Overall Goal, Criteria, and Decision Alternatives • Step 2: Develop a Pair-wise Comparison Matrix (Criteria) Scales for the importance are (1/9, 1/8, . . . 1, 2, . . ., 9) Perform a consistency test • Step 3: Develop a Normalized Matrix and the Priority Vector (Criteria) • Step 4: Develop a Pairwise Comparison Matrices of alternatives for each criterion (Alternatives) Scales are (1/9, 1/8, . . . 1, 2, . . ., 9) Perform consistency tests • Step 5: Devlop a Normalized Matrix and the Priority Vector (Alternatives) • Step 6: Get the weighted average using s j wi rij i • Step 7: Choose the best alternative © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 33 Analytic Hierarchy Process How to handle subcriteria Perform Steps 2~6 for subcriteria. Perform Steps 2 ~ 7 for higher level criteria. (see slide 24) Consistency Test • Are we sure that pairwise comparison matrix is consistent? If A >> B(A is preferred to B) and B >> C, then A >> C • Perfect consistency a(i, k) = a(i, j) x a(j, k) for all i, j ex. 6 = 2 x 3 (C : B : A) = 1 : 3 : 6 see Excel Sheet If the pairwise comparison matrix is perfectly consistent, the largest eigen value max n • Number of perfect consistent matrix of n = 3 is 43. (1, 1, 1), (1, 2, 2), . . . (1, 9, 9), (2, 2, 4), (2, 3, 6), (2, 4, 8) (3, 3, 9) etc. (8+2)x2x2 + 3 = 43 out of 19^3 = 6859 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 34 Analytic Hierarchy Process Consistency Test • Calculate max (p.677) Multiply each column of comparison matrix with the corresponding element of priority vector Get the row sum of the matrix obtained from above Divide each element of row sum vector (obtained above) by the corresponding element of priority vector max is the average of elements of the vector obtained above • Consistency Index CI max n n 1 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 35 Analytic Hierarchy Process Consistency Test max n CI • Consistency Ratio n 1 CI CR RI Where RI is the value depending on the number of criteria or alternatives. n RI 3 4 5 6 7 8 9 10 11 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49 1.51 If CR <= 0.1, then the pairwise comparison matrix is considered as consistent. •see Excel Sheet © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 36 End of Chapter 14 © 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a © 2008 Thomson South-Western. All Rights Reserved publicly accessible website, in whole or in part. Slide 37