Hi, I can see this Salesmen Travelling Problem is symmetric, i.e., the distance matrix is symmetrical, the distance traveled between two points is the same. By the same way, can I proposed the pheromone travelled between two points in this problem be the same? i.e, the pheromone matrix is symmetrical in the Salesmen Travelling Problem? Will this through affect the solution?
Prof Lam, At 11:50 for the AS method: why the contributions from ant 1, 2 and 5 are 1/15? Would you please elaborate on this? Should they be the same as SACO (1/35, 1/55, 1/40)? Thank you.
Thanks for your comment. The contributions from 1, 2 and 5 are 1/15 because they follow the Ant-quantity AS contribution rule given on the bottom right-hand side of the slide at 11:50. The update rule is Q/d_{ij}(t).
Great video. It helped a lot to implement the algorithm in the context of machine scheduling.
Very helpful!
Super helpful video, thanks heaps : )
thank you sir
Sir please explain, if the original path chosen for each ant in the beginning is through transition probabilities and if yes how, or if it is random
Hi, I can see this Salesmen Travelling Problem is symmetric, i.e., the distance matrix is symmetrical, the distance traveled between two points is the same.
By the same way, can I proposed the pheromone travelled between two points in this problem be the same? i.e, the pheromone matrix is symmetrical in the Salesmen Travelling Problem?
Will this through affect the solution?
Thank you professor
Thank you.
@@hklam2368 Can i contact you regarding this presentation there is point i didn't understand and I'm working on a AI project? thanks
always nice to learn a thing or two from elon musk.
Prof Lam, At 11:50 for the AS method: why the contributions from ant 1, 2 and 5 are 1/15? Would you please elaborate on this? Should they be the same as SACO (1/35, 1/55, 1/40)? Thank you.
Thanks for your comment. The contributions from 1, 2 and 5 are 1/15 because they follow the Ant-quantity AS contribution rule given on the bottom right-hand side of the slide at 11:50. The update rule is Q/d_{ij}(t).
支持