I need to figure out the total path (A to Z) followed by an agent through a squared-element grid. Each grid element has a probability density function Delta assigned to it that represents the probable direction (0-360) that the agent might take when stepping over it.
So, say that when agent is in grid cell A, I sample the corresponding directions pdf Delta_A and I get the value delta_A. Agent then moves in that direction delta and arrives to a neighbor grid cell in the next step B.
To find out next direction, I'll sample the pdf Delta_B, but here I want the value delta_B to receive more weight than the other values in Delta_B. Is there a methodology for that?
Thanks!
