Spreaditround Ltd., is a large fertilizer company with offices scattered throughout the United States. A communications network links the various offices to the head office in Detroit. The offices have online real-time update capabilities to several centralized databases.

Response times in the network have been deteriorating. After an investigation into the possible reasons for the increase in response times, the system programming group suggests two alterations to the network that they believe could remedy the situation. The first is to purchase a new model of network controller. They are uncertain, however, as to which of two models of controller to purchase. The second option is to change the method of polling terminals in the system from roll-call to hub polling.

The system programming group constructs a simulation model of the network and runs the model for 2,000 iterations. The first 1,000 iterations allow the simulation to stabilize. During the second 1,000 iterations, response times for a particular controller-method of polling configuration are measured. The experiment is run four times, one for each controller-method of polling combination.

As the manager of internal audit for Spreaditround, you receive a report from the system programming group on their simulation runs that contains the following table:

Required: On the basis of the simulation results, what do you conclude about the proposed changes? If a statistical analysis (ANOVA) of the results were undertaken, what variables do you think would be statistically signifcant? What kinds of concerns might you have about the validity of the results obtained using the simulation model?

Transcribed Image Text:

Mean Response Times in Seconds for 1,000 Iterations of Each Controller-Method of Polling Combination Type of Controller Method of Polling Type 1 Type 2 Row Mean Roll-call 5.0020 4.1570 4.5794 Hub 5.0010 3.0620 4.0315 Column Mean 5.0015 3.6095 4.3055