Search strategy in a Random Grid

Suppose we have a MxN grid with each cell having a particular (random) number. Also suppose that you are an organism capable of moving one cell left/right/up/down and accessing the number on the cell you are presently standing on. Moving takes away X units of energy/time and accessing the number takes Y units/time. You are assigned the task to search for a particular number NUM in the grid using the least amount of energy/time. You are allowed to choose the location (middle/corner/random) where you want to start searching. What will be your search strategy?

Will you do a systematic search or a clumsy search? Well, this is what I want to find out. I want to find out which is the best search strategy when there is a random chance of hitting upon the goal. Maybe, someone ( me, perhaps! ) should apply genetic programming to evolve agents who do this and then analyze their strategy.

This problem can be very exciting and can have huge applications.

3 comments

  1. I dare to say:
    if the gird is totally randomized, there is not much you can do, except to make sure that you choose a path where you traverse and access every cell only once.
    if the gird is not 100% random, it depends on how exactly the values of the cells relate to each other, and there might be an ideal algorithm.

    can you agree?

  2. Assuming that you have N algorithms (A1 to An), and two different, totally random grids X and Y.

    Maybe A1 will be best for X, and A2 best for Y. I see no possibility to say anything in advance if there is only random data. But if there is the slightest bit of non-random in your grid, and if you KNOW something about it, then there is a chance for on optimal algorithm that takes this information into account.

Comments are closed.