Multiple traveling salesmen how to calculate possible routes (VRP with Timewindows)

[ev13wt]

Do we have a math guru here?

For a single traveling salesman, the calculation is easy: (N-1)! / 2
/2 becaues X,A,B,C,XX,A,B,C,X; X,C,B,A,XX,C,B,A,X are considered as "the same"

Now, I have 5 Technicians that must visit 10 customers each. Above is 1 tech visiting.
How do I figure out this? I need to prove a machine is better than a human.


If there are 12 cities to visit, how many possible routes are?
Are there (11*10*9*8*7*6*5*4*3*2*1)/2 = 19,958,400 routes?



Disregarding time windows:
Number of techs to exponent (N-1)!/2 ?
I'm lost - brain freeze.


https://en.wikipedia.org/wiki/Vehicle_routing_problem
https://en.wikipedia.org/wiki/Travel...lesman_problem

 

[plym49v2]

You don't want routes, you want optimal routes. What you ae asking for is the optimal integer solution for the Travelling Salesman problem, and with 10 salespeople to boot. I don't think that is solveable.

 

[Jackie Treehorn]

We’re not going to do your homework for you, son.

 

[ev13wt]

You don't want routes, you want optimal routes. What you ae asking for is the optimal integer solution for the Travelling Salesman problem, and with 10 salespeople to boot. I don't think that is solveable.

We already have the technology and the algorithm. The thing optimizes 25 techs with 10 orders each, with time window constraints, while adhering to "skills needed" in about 2 seconds flat right now. Saving 30% time and miles already. But it seems the human is the problem.

I am trying to sell it, and everybody planners all say: "Shit, I know my way around the xxxxx area much better than any computer."
I need to calculate a figure of how many possibilites that is, to prove they are wrong and that it is fucking impossible, nobody can calculate more than 1 tech and 5 orders. PLS.

And I need to show my calculation - it needs to be credible. Else I'd just take amountTechs to the power of amountOfOrders and write some random shit number.

 

[Ted_Kaczynski]

We already have the technology and the algorithm. The thing optimizes 25 techs with 10 orders each, with time window constraints, while adhering to "skills needed" in about 2 seconds flat right now. Saving 30% time and miles already. But it seems the human is the problem.

I am trying to sell it, and everybody planners all say: "Shit, I know my way around the xxxxx area much better than any computer."
I need to calculate a figure of how many possibilites that is, to prove they are wrong and that it is fucking impossible, nobody can calculate more than 1 tech and 5 orders. PLS.

And I need to show my calculation - it needs to be credible. Else I'd just take amountTechs to the power of amountOfOrders and write some random shit number.

If you're selling it, you should already know the answer BEFORE the attempt to pawn it off. Seriously, you're coming here for this?

Gonna share your paycheck with the user here that helps you out?

 

[ev13wt]

I've seen micromanagement before but holy shit dude...

If you see the scale it makes a lot of sense.
Our tech saves our customers hard cash.

Before optimization: 2900 miles
After: 1800 miles and a full tech is empty and has nothing to do.

It's a shit ton of money, our customers have about 300 to 800 techs each. (20 customers)

 

[ev13wt]

If you're selling it, you should already know the answer BEFORE the attempt to pawn it off. Seriously, you're coming here for this?

Gonna share your paycheck with the user here that helps you out?

I suck at math.
Maybe someone here likes math.
It's an interesting question for some.

The end.

 

[Romogo]

If you're selling it, you should already know the answer BEFORE the attempt to pawn it off. Seriously, you're coming here for this?

Gonna share your paycheck with the user here that helps you out?

No, they will just be happy to do their part Comrade.

 

[ExWrench]

Multiple traveling salesmen how to calculate possible routes

Good luck identifying your father among the vast pool of candidates

 

[m016324]

If you have the software it shouldn't be that hard. Have the routers put together their routes. Have the system optimize routes. Compare mileage and manhours between the two. Did this for hundreds of optimization projects in logistics. Only had one router that could consistently beat the computer. He had 30 years of experience in a some what unique area (tractor trailer DSD) and he could double load trailers that the routing software couldn't figure out. From there it's just a labor and cost per mile calculation. No need to show the work with all the mapping software that exists.

-ben

 

[Jackie Treehorn]

Good luck identifying your father among the vast pool of candidates

If Mike goes out for smokes at 8:15 on a Tuesday and his ‘82 el Camino is traveling at 38 mph, how many years does it take to send eviltwit a birthday card?