Author: h.g.muller
Date: 00:55:23 02/11/06
Go up one level in this thread
This whole contest seems malicious and can be described as 'cruel and unusual punishment'. The Zombi problem seems even worse than the chess problem, I wouldn't know how to solve it other than by a horrendous tree-search, that would explode for all but the smallest office buildings. It seems related to the traveling-salesman problem, and the best hope to write a program that sometimes produces a result, rather than to hang up on you, would be to use a heuristic to find an approximate solution. (e.g. evolutionary improving on a path between the guard and the employee). The chess problem at least was limited in computational-resourch needs because it was guaranteed to be mate in one, on a board of defined size... For the cone-sphere problem the mathematics is far from trivial: there is no obvious best position of the sphere from geometric arguments, finding the solution requires calculus to optimise a cubic function. There is no easy numerical trial-and-error solution. The required precision rules out the use of Monte-Carlo techniques to calculate the volume, you would need millions of points to get the statistical noise down to that level even for a given position and radius of the sphere. And even in that case you would have to be aware of the equations of analytical geometry describing cone and sphere surfaces, to determine which points ly inside or outside. Even a wizzard in programming does not stand a chance on this problem if he lacks advanced mathematical knowledge. The chess-problem only seems at the edge of do-able because we are all experts in that area with lots of experience. We know the tricks of how to delimit the board and recursive search, and what is involved in testing for checkmate. I am not surprised that only solutions to the magic-square problem where turned in!
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.