|
| 1 | +a. Fred's 20-minute wait does not change the distribution of the additional waiting time for the next bus, due to the memoryless property of the exponential. |
| 2 | +Since the intervals between Route 1 buses are i.i.d. $\mathrm{Expo}(\lambda_1)$, Fred will have to wait another $1/\lambda_1$ minutes on average. \\ |
| 3 | + |
| 4 | + |
| 5 | +b. Let's define the r.v.s of interest below |
| 6 | + |
| 7 | +\begin{itemize} |
| 8 | +\item $T_1$: time of arrival of the first Route 1 bus; |
| 9 | + |
| 10 | +\item $\Delta_i$: time interval between the i-th and (i+1)-th arrivals of Route 1 buses; |
| 11 | + |
| 12 | +\item $T_2$: time of arrival of the first Route 2 bus; |
| 13 | + |
| 14 | +\item $B$: number of Route 1 bus arrivals before the first Route 2 bus. |
| 15 | +\end{itemize} |
| 16 | + |
| 17 | +By the property of the Poisson process, $T_1 \sim \mathrm{Expo}(\lambda_1)$ and\\ $T_2 \sim \mathrm{Expo}(\lambda_2)$. |
| 18 | + |
| 19 | +The event $\{B=k\}$ happens when each of the first k Route 1 buses arrives before $T_2$, and the (k+1)-th Route 1 bus arrives after $T_2$. |
| 20 | +These sub-events are independent due to the memoryless property of the exponential. |
| 21 | + |
| 22 | +The first Route 1 bus arrives before the Route 2 bus with probability |
| 23 | +$P(T_1 < T_2) = \lambda_1/(\lambda_1+\lambda_2)$. |
| 24 | +The probability that the second Route 1 bus arrives before the Route 2 bus is $P(\Delta_1 < T_2 - T_1)$; once $T_1$ realizes to a fixed value, $T_2-T_1$ represents the additional waiting time for the Route 2 bus after the arrival of the first Route 1 bus, thus $T_2-T_1 \sim \mathrm{Expo}(\lambda_2)$ by the memoryless property. Therefore, $P(\Delta_1 < T_2 - T_1) = \lambda_1/(\lambda_1+\lambda_2)$. |
| 25 | +With the same reasoning, we conclude that the probability for each of the first k buses is given by the same expression. |
| 26 | +Finally, the (k+1)-th Route 1 bus arrives after $T_2$ with probability $P(\Delta_k > T_2 - \sum_{i=1}^{k-1} \Delta_i) = \lambda_2/(\lambda_1+\lambda_2)$. |
| 27 | +Therefore, |
| 28 | + |
| 29 | +$$ |
| 30 | +P(B=k) = \left( \frac{\lambda_1}{\lambda_1+\lambda_2} \right)^k \left( \frac{\lambda_2}{\lambda_1+\lambda_2} \right) \text{ , for } k = 0, 1, \dots |
| 31 | +$$ |
| 32 | + |
| 33 | +The event of interest is $\{B \ge n\}$, calculated as |
| 34 | + |
| 35 | +\begin{flalign*} |
| 36 | +P(B \ge n) |
| 37 | +& = \sum_{k=n}^\infty P(B=k) = \frac{\lambda_2}{\lambda_1+\lambda_2} \sum_{k=n}^\infty \left( \frac{\lambda_1}{\lambda_1+\lambda_2} \right)^k \\ |
| 38 | +& = \left( \frac{\lambda_1}{\lambda_1+\lambda_2} \right)^n |
| 39 | +\end{flalign*} |
| 40 | + |
| 41 | + |
| 42 | +c. Let $W_F$ and $W_G$ be Fred's and Gretchen's waiting times, respectively. |
| 43 | +$W_F$ and $W_G$ are i.i.d. $\mathrm{Expo}(\lambda)$. |
| 44 | + |
| 45 | +The first of the two, Fred or Gretchen, takes the bus at \\ |
| 46 | +$L = \min(W_F,W_G) \sim \mathrm{Expo}(2\lambda)$. |
| 47 | +After that, the second of the two waits an additional time $t \sim \mathrm{Expo}(\lambda)$, by the memoryless property. |
| 48 | + |
| 49 | +The time it takes until both Fred and Gretchen have caught their buses is |
| 50 | +$M = L + t$. |
| 51 | +Taking the expected value |
| 52 | + |
| 53 | +$$ |
| 54 | +E(M) = E(L) + E(t) = \frac{1}{2\lambda} + \frac{1}{\lambda} = \frac{3}{2\lambda} |
| 55 | +$$ |
0 commit comments