Hey everyone, I recently stumbled upon NN and I was wondering is it possible using NN to create a software that could handle logistic operations of a shipping company. Inputs would be sets of variables such as available routes, trucks, deliveries, employees and the outputs would be an organized schedule for the next day (for example). Is this feasible using NN? Thanks for your time, Adrian

Generally speaking, neural networks recognise patterns, and that's not really what you want here. If you want a canned technology for this task, then genetic algorithms are known to be quite good at scheduling problems.

Quote: Original post by Kylotan
Generally speaking, neural networks recognise patterns, and that's not really what you want here. If you want a canned technology for this task, then genetic algorithms are known to be quite good at scheduling problems.

Thanks, I appreciate your support.

Adrian

Also, consider using a non-AI approach, like PERT.

Or you could use any of the constraint programming methods...such as MILP (Mixed Integer Linear Programming).

There is a HUGE amount of literature in Operations Research about how to solve problems of the kind you have. Grab your mouse and start Googling! ;)

Cheers,

Timkin