Results (
Indonesian) 1:
[Copy]Copied!
5.1 OVERVIEWThe Integer Linear Programming module of The Management Scientist employs a branch and bound solution procedure. It can solve all-integer, mixed-integer, and 0–1 integer linear programs with up to 100 variables and 50 constraints. However, problems with large numbers of integer variables may solve slowly. Just as with the Linear Programming module, the number of deci¬sion variables that can be used must allow for slack, surplus, and artificial variables that are added automatically by the Integer Linear Programming module. Since the Linear Programming module and the Integer Linear Program¬ming module are the same with respect to creating, retrieving, saving, and editing a problem, we will focus our attention on the differences associated with solving integer linear programs. Readers unfamiliar with the Linear Pro¬gramming Module of The Management Scientist should review Chapter 2 before continuing with this chapter.5.2 AN EXAMPLE PROBLEM: THE ALL-INTEGER CASEEastborne Realty currently has $2,000,000 available to purchase town¬houses or apartment buildings. Each townhouse can be purchased for the price of $282,000, but there are only five town¬houses available for purchase at this time. Each apartment building sells for $400,000, and the developer has agreed to build as many units as Eastborne would like to purchase. Eastborne’s property manager is free to devote 140 hours per month to these investments. Each townhouse will require 4 hours of the prop¬erty manager’s time each month, while each apartment building will require 40 hours per month. The yearly cash flow (after deducting mortgage pay¬ments and operating expenses) is estimated at $10,000 per townhouse and $15,000 per apartment building. Eastborne would like to allocate its invest¬ment funds to townhouses and apartment buildings in order to maximize the yearly cash flow. To develop an appropriate mathematical model, let us introduce the fol¬lowing decision variables: T = number of townhouses purchased A = number of apartment buildings purchasedThe objective function, measuring cash flow in thousands of dollars, can be written as max 10T + 15AThere are three constraints that must be satisfied: 282T + 400A 2000 Funds available ($1000s) 4T + 40A 140 Manager’s time (hours) T 5 Townhouses availableAdding the nonnegativity requirements we obtain a linear programming model involving the two variables and three constraints. This model could be solved using the Linear Programming module described in Chapter 2; the optimal linear programming solution is T = 2.479 and A = 3.252, with an objective function value of $73,574. However, since fractional values do not make any sense in the context of the problem, the standard linear programming approach to this problem is not appropriate. What we need to do is add the requirement that both T and A are restricted to integer values. Doing this, we obtain the following all-integer linear programming model: max 10T + 15A s.t. 282T + 400A 2000 Funds available ($1000s) 4T + 4A 140 Manager’s time (hours) T 5 Townhouses available T, A 0 and integer5.3 CREATING AND SOLVING A PROBLEMTo solve the Eastborne Realty problem, we begin by selecting the Integer Linear Programming module and choosing New from the File menu; the Problem Features dialog box will then appear. Figure 5.1 shows the Problem Features Dialog box after entering 2 for the Number of Decision Variables, 3 for the Number of Constraints, and choosing Maximize for the Optimization Type. After clicking OK, we obtain the data input screen shown in Figure 5.2. Note that the two decision variables have initially been given the names X1 and X2, and that the three constraints are labeled Constraint 1, Constraint 2, and Constraint 3.
Being translated, please wait..
