The document describes a capacity planning problem for Communicorp using integer programming to determine the optimal server configuration over multiple periods. It presents three models:
1) A monthly analysis model that examines each month individually and finds the optimal servers for each.
2) A two-phase analysis that considers setup and operating costs over two periods and constraints like budget and capacity.
3) A two-period model that also incorporates server life expectancies and constraints to ensure servers from the first period are replaced in the second if needed. The three-sentence summary compares the optimal costs of each model.
11. Model 3 – Two Period Life
Expectancy
SERVER TYPE LIFE EXPECTANCY
IBM – X1 5
Standard Intel - X2 2
Enhanced Intel – X3 3
SGI workstation – X4 3
Sun Workstations – X5 4
Lenovo – X6 3
Dell- X7 3
X1-x7: BINARY VARIABLESX8-X14: GENERAL INTEGER VARIABLES
MONTH 2: CHANGE IN OBJECTIVE FUNCTIONUNIT COST OF SGI WORKSTATION = 10% DISCOUNT = X11 CHANGES FROM 10,000 TO 9,000UNIT COST OF SUN WORKSTATION = 25% DISCOUNT = X12 CHANGES FROM 25,000 TO 18,750CONSTRAINT:- CAPACITY - SET-UP COST INCURRED WHEN A SERVER IS PURCHASED: The coefficient in front of the binary variable (setup cost) represent the Big M technique where the Big M carries the value of 100 000. In the case that binary variable for the set up cost is equal to 1; we want to ensure the model purchases that type of server. Subsequently, if the value of the binary variable for the set up cost is zero due to the minimization nature the model, the model will not incur in the variable cost giving the general integer variable the value of zero. X8-100000X1 <=0