Yes, very interesting. I rather recoil at the thought of human behaviour being reduced to a computer algorithm. Though in a lost/survival situation, we basically run that sort of algorithm in our heads. Only with more random factors, like walking in circles.

Given all the information already in GIS databases, you could run this software to get a rapid, theoretical list of "hot spots" for a person's location at a certain time frame. Could save lives and a lot of time and money too. If it really works.

It would be a wonderful SAR training tool. Input a true scenario and have search managers pit their experience against the model's predictions. Then compare both to the actual outcome. That would be very interesting indeed.