If you had a traveling salesman problem with 1000 points, what's

greedy

>>7663110
>what's the first thing you would do to solve it?
I would solve it.

>>7663158
/thread

>>7663110
I wrote an ant colony optimization simulation for """solving""" this problem for my java class last year.

>>7663183
How close did it get to the optimal solution?

>>7663110
visiting wiki and looking up solving algorithms

>>7663154
That's awful for TSP tho senpai.

some sort of heuristic

>>7663110
You should probably research known approximation algorithms for this problem.

>>7663110
deep recurrent neural network with genetic step function

try generating lots of random walks and hope that statistics of the graph are nice enough

>>7663110
You don't solve it you approximate it. This can literally range from greedy as suggested before to a whole range of more complex shit. Basically if you need to show you've done something do a greedy and then make it better with 2-opt.

>>7663244
I don't know what the optimal solution was, but the solutions I got seemed pretty good.

>>7664139
still the first thing I would try
if only to see if my boilerplate works