| You're playing a video game that features **N** different areas, numbered from | |
| 1 to **N**. There are **M** one-way paths that each connect two areas. The | |
| **i**th path runs from area **Ai** to a different area **Bi**. No pair of | |
| paths directly connect the same pair of areas, and for every area it's | |
| impossible to start at that area and follow a non-empty sequence of paths to | |
| return to that area. In other words, the game's layout is a directed acyclic | |
| graph. | |
| You start in area 1. **K** other distinct areas each contain an item to | |
| collect — the **i**th of these is area **Ii**. As soon as you've visited these | |
| **K** areas at least once each, you win! You'd like to do so as fast as | |
| possible. | |
| At any point in time, if there are no outgoing paths leading away from your | |
| current area, you automatically respawn in area 1 after a delay of **R** | |
| seconds. Otherwise, you get to choose one such path and attempt to follow it. | |
| Unfortunately, this game relies entirely on Random Number Generation to | |
| determine whether or not you'll be successful, regardless of your skill. In | |
| particular, with probability **P**, you'll successfully travel along your | |
| chosen path for **D** seconds, ending up in a new area. On the other hand, | |
| with probability 1 - **P**, you'll instead perish and respawn in area 1 after | |
| a delay of **R** seconds. | |
| What's the minimum expected time for you to collect all **K** items, given | |
| that you play optimally? This is guaranteed to be possible — that is, all | |
| **K** areas that contain items are reachable from area 1. Your output should | |
| have at most 10-6 absolute or relative error. | |
| ### Input | |
| Input begins with an integer **T**, the number of times you play the game. For | |
| each time, there is first a line containing the space-separated integers | |
| **N**, **M**, and **K**. The second line contains the space-separated integers | |
| **D** and **R**. The third line contains the real value **P** which is given | |
| with at most 4 decimal places. The fourth line contains the **K** space- | |
| separated integers **I1** to **IK**. Then, **M** lines follow, the **i**th of | |
| which contains the space-separated integers **Ai** and **Bi**. | |
| ### Output | |
| For the **i**th time you play the game, print a line containing "Case #**i**: | |
| " followed by the expected time it will take you to collect all of the items | |
| if you play optimally. | |
| ### Constraints | |
| 1 ≤ **T** ≤ 50 | |
| 2 ≤ **N** ≤ 100,000 | |
| 1 ≤ **M** ≤ 100,000 | |
| 1 ≤ **K** ≤ min(20, **N** \- 1) | |
| 0.5 ≤ **P** ≤ 1.0 | |
| 1 ≤ **D**, **R** ≤ 1,000 | |
| 1 ≤ **Ai**, **Bi** ≤ **N** | |
| 2 ≤ **Ii** ≤ **N** | |
| The answer for each game is guaranteed to be less than 1030. | |
| ### Explanation of Sample | |
| In the first game, it takes you 10 seconds to reach the only item, and then | |
| you win. There's no chance of failure. The second game is the same as the | |
| first, but now you fail to reach the item with probability 0.5. On average you | |
| will fail once before reaching the item, so you'll incur an average penalty of | |
| 3 seconds on top of the 10 seconds it takes you to succeed, for a total of 13 | |
| seconds. | |