Consider the two systems of Example 10.1, with the same initial conditions and performance...

Consider the two systems of Example 10.1, with the same initial conditions and performance measures given there; let ζ = d_Z(100) – d_K(100).

(a) Make n₁ = n₂ = 5 independent replications of each system and construct an approximate 90 percent confidence interval for ζ. (Perform the simulations for the two systems independently of each other.) Use the paired-t approach.

---

(b) Make n₁ = 5 replications of the Zippytel system and n₂ = 10 replications of the Klunkytel system, and construct an approximate 90 percent confidence interval for ζ. (Again make the runs of the two systems independently.)

---

(c) Use the procedure of Sec. 10.4.1 to select the best of the k = 2 systems. Let n₀ = 20, P* = 0.90, and d* = 0.4.

Example 10.1

FIGURE 10.1 One Zippy or two Klunkys?