14
Linear Programming Models in Services

DocumentLP

Embed Size (px)

DESCRIPTION

LP in services

Citation preview

  • Linear Programming Models in Services

  • Learning ObjectivesDescribe the features of constrained optimization models.Formulate LP models for computer solution.Solve two-variable models using graphics.Explain the nature of sensitivity analysis.Solve LP models with Excel Add-in Solver and interpret the results.Formulate a goal programming model.

  • Stereo WarehouseLet x = number of receivers to stock y = number of speakers to stockMaximize 50x + 20y gross profitSubject to 2x + 4y 400 floor space 100x + 50y 8000 budget x 60 sales limit x, y 0

  • Diet ProblemLakeview Hospital Let E = units of egg custard base in the shake C = units of ice cream in the shake S = units of butterscotch syrup in the shake

    Minimize Subject to cholesterol fat protein calories

  • Shift-Scheduling ProblemGotham City Police Patrol Let xi = number of officers reporting at period i for i =1, 2, 3, 4, 5, 6

    Minimize x1 + x6 6 period 1 x1 + x2 4 period 2 x2 + x3 14 period 3 x3 + x4 8 period 4 x4 + x5 12 period 5 x5 + x6 16 period 6

  • Workforce-Planning ProblemLast National Drive-in Bank Let Tt = number of trainees hired at the beginning of period t for t = 1,2,3,4,5,6 At = number of tellers available at the beginning of period t for t = 1,2,3,4,5,6 Minimize subject to

    A1 = 12 for t = 2,3,4,5,6 At , Tt 0 and integer for t = 1,2,3,4,5,6

  • Transportation ProblemLease-a-Lemon Car Rental Let xij = number of cars sent from city i to city j for i = 1,2,3 and j = 1,2,3,4

    Minimize 439x11 + 396 x12 + . . . +479x33 + 0x34 subject to x11 + x12 + x13 + x14 = 26 x21 + x22 + x23 + x24 = 43 x31 + x32 + x33 + x34 = 31 x11 + x21 + x31 = 32 x12 + x22 + x32 = 28 x13 + x23 + x33 = 26 x14 + x24 + x34 = 14 xij 0 for all i , j

  • Graphical SolutionStereo Warehouse Z=2000Z=3000Z=3600Z=3800ABCDEOptimal solution( x = 60, y = 40)

  • Model in Standard FormLet s1 = square feet of floor space not used s2 = dollars of budget not allocated s3 = number of receivers that could have been sold

    Maximize Z = 50x + 20ysubject to 2x + 4y + s1 = 400 (constraint 1) 100x + 50y + s2 = 8000 (constraint 2) x + s3 = 60 ( constraint 3) x, y, s1, s2, s3 0

  • Stereo WarehouseExtreme-Point SolutionsExtreme Nonbasic Basic Variable Objective-function point variables variables value value Z A x, y s1 400 0 s2 8000 s3 60 B s3, y s1 280 3000 s2 2000 x 60 C s3, s2 s1 120 3800 y 40 x 60 D s1, s2 s3 20 3600 y 80 x 40 E s1, x s3 60 2000 y 100 s2 3000

  • Sensitivity AnalysisObjective-Function Coefficients z = 50x + 20y (constraint 3 )(constraint 1)(constraint 2)ABCD

  • Sensitivity AnalysisRight-Hand-Side Ranging (constraint 3 )(constraint 2)ABICDH

  • Goal ProgrammingStereo Warehouse ExampleLet x = number of receivers to stock y = number of speakers to stock = amount by which profit falls short of $99,999= amount by which profit exceeds $99,999= amount by which floor space used falls short of 400 square feet= amount by which floor space used exceeds 400 square feet= amount by which budget falls short of $8000= amount by which budget exceeds $8000= amount by which sales of receivers fall short of 60= amount by which sales of receivers exceed 60= priority level with rank kMinimizesubject to profit goal floor-space goalbudget goal sales-limit goal

  • Topics for DiscussionHow can the validity of LP models be evaluated?Interpret the meaning of the opportunity cost for a nonbasic decision variable that did not appear in the LP solution.Explain graphically what has happened when a degenerate solution occurs in an LP problem.Is LP a special case of goal programming? Explain.What are some limitations to the use of LP?