Bessie, always wishing to optimize her life, has realized that she really enjoys visiting F (1 <= F <= P) favorite pastures Fi of the P (1 <= P <= 500; 1 <= Fi <= P) total pastures (conveniently
numbered 1..P) that compose Farmer John's holdings.
Bessie knows that she can navigate the C (1 <= C <= 8,000) bidirectional cowpaths (conveniently numbered 1..C) that connect various pastures to travel to any pasture on the entire farm. Associated with each path Pi is a time Ti (1 <= Ti <= 892) to traverse that path (in either direction) and two path endpoints ai and bi (1 <= ai <= P; 1 <= bi <= P).
Bessie wants to find the number of the best pasture to sleep in so that when she awakes, the average time to travel to any of her F favorite pastures is minimized.
By way of example, consider a farm laid out as the map below shows, where *'d pasture numbers are favorites. The bracketed numbers are times to traverse the cowpaths.
1*--[4]--2--[2]--3
| |
[3] [4]
| |
4--[3]--5--[1]---6---[6]---7--[7]--8*
| | | |
[3] [2] [1] [3]
| | | |
13* 9--[3]--10*--[1]--11*--[3]--12*
The following table shows distances for potential 'best place' of pastures 4, 5, 6, 7, 9, 10, 11, and 12:
* * * * * * Favorites * * * * * *
Potential Pasture Pasture Pasture Pasture Pasture Pasture Average
Best Pasture 1 8 10 11 12 13 Distance
------------ -- -- -- -- -- -- -----------
4 7 16 5 6 9 3 46/6 = 7.67
5 10 13 2 3 6 6 40/6 = 6.67
6 11 12 1 2 5 7 38/6 = 6.33
7 16 7 4 3 6 12 48/6 = 8.00
9 12 14 3 4 7 8 48/6 = 8.00
10 12 11 0 1 4 8 36/6 = 6.00 ** BEST
11 13 10 1 0 3 9 36/6 = 6.00
12 16 13 4 3 0 12 48/6 = 8.00
Thus, presuming these choices were the best ones (a program would have to check all of them somehow), the best place to sleep is pasture 10.