that was included in Exercise 4-25), but expand the output performance measures to the mean total time in system of parts sorted according to each of:

(a) Part As, regardless of how/where they exit (Shipped, Salvaged, or Scrapped)

(b) Part Bs, regardless of how/where they exit

(c) All parts together, regardless of type (A or B) and how/where they exit

(d) Parts exiting via Shipped, regardless of part type

(e) Parts exiting via Salvaged, regardless of part type

(f) Parts exiting via Scrapped, regardless of part type Modify your models as needed to get all six of these output performance measures. In addition to just running both models with this expanded set of output performance measures, use the Statistic Data Module, and then the Arena Output Analyzer to do a paired t test (and conf dence intervals) of the means in the subtraction direction (With AJ) – (Without AJ); according to these results, for which output performance measures does the presence of AJ make a statistically signif cant difference on the means? Also, use the Process Analyzer (PAN) to run the two scenarios together in a single PAN run (you’ll need different .p Program Files for each scenario) and make six box-and-whisker plots, one for each of the six output performance measure and each plot containing both the “with-AJ” and the “without-AJ” results, and use PAN’s selection algorithm (smaller values on average total time in system are better, of course) to identify, for each of the six responses, whether one is statistically signif cantly better than the other in terms of means. Do your PAN results and your Output Analyzer results agree with each other? Supply a brief write-up of your f ndings, including screen shots, tables, or plots as you see f t to make your points.