At UPS, the Algorithm Is the Driver

[cosmo1]

Our dispatcher is the red headed step child bastard of EDD. He doesn't care how bad his dispatches are. Our DOLs have been jacked up since day one (going on nine years now) he refuses to fix anything. And now they expect ORION to work? Now that's funny.

If I didn't know better, I'd say you work in my old building.

 

[Brownslave688]

If I didn't know better, I'd say you work in my old building.

And mine.

 

[Jones]

My last PDS in delivery was the absolute worst but really I owe him one because if it wasn't for him I might never have signed that feeder bid sheet.

 

[BrownArmy]

We used to have a superhero for dispatch, but he was transferred to another building.

The new guy just doesn't give a chit, and it shows.

Just terrible, really really terrible.

Half the routes in my center have good DOLs, the rest are complete garbage.

It's like we need an act of Congress to make any changes, so you know, garbage in, garbage out.

 

[Signature Only]

TSP = Total Service Plan? Care to break it down for us?

For a detailed explanation just google TSP or see Wikipeda.

I'll give you the basics. TSP (Traveling Salesman Problem) on which Orion is based originated in the 1800s as Irish mathematician William R. Hamilton considered a way to compute the shortest route a salesman could take upon leaving his hometown, calling on his accounts and returning home.

Later British mathematician Thomas Hinkman's work led to Karl Menger proposing a brute force algorithm (computing every possible route and selecting the one with the shortest distance).

Around 1930 Princeton mathematician Hassler Whitney named this concept TSP.

Considerable contributions in refining algorithms where made by George Dantzig, Selmer Johnson and the renown Delbert Fulkerson at Santa Monica's Rand Corporation in the 1960s.

Until this point advancements in computers and algorithms led to more and more cities or in our case, stops or individual addresses being considered for solution inclusion.

Dr. Arora broke ground when he found or rather envisioned a way to restate the TSP concept as a calculus equation to which the first branch of calculus (the derivative) would be useful.

Upon solving the equation he found that as friend(x) where x=optimal solution, approaches infinity (considering the many variables and with a permissible solution acquisition time of 1 hour) the result was 85%.

This meant that regardless of available computational power or algorithm elegance, the best result that can ever be achieved is 85%.

You should look deeper into this particularly in the field of cognitive research as scientist there have shown repeatedly that humans with no previous training can outperform the best computer and best algorithms, arriving at solutions 92 to 97 percent of optimal within 5 minutes.

The best computers would take a month to match that.

And computers don't provide an 85% result constantly. Remember 85% is the very best you'll ever get. Most times its between 30 to 60% of optimal. Just change 1 thing in a DOL and rerun Orion and see how many different results you get. Then consider the nearly infinite variables a driver takes into account each and every day.

So now, cognitive research is centered on discovering what enables humans to do such a good job with so many variables in such a short amount of time. If an algorithm could be formulated that would be half as good imagine the applications.

The defense industry has.

By the way, Dr. Arora's work has been authenticated by the mathematical and computer community. Hence the Godel Prize in 2008.


A bit long winded, I know. But check his research and you'll see that Orion is nothing but a blind alley.

 

[TBH]

For a detailed explanation just google TSP or see Wikipeda.

I'll give you the basics. TSP (Traveling Salesman Problem) on which Orion is based originated in the 1800s as Irish mathematician William R. Hamilton considered a way to compute the shortest route a salesman could take upon leaving his hometown, calling on his accounts and returning home.

Later British mathematician Thomas Hinkman's work led to Karl Menger proposing a brute force algorithm (computing every possible route and selecting the one with the shortest distance).

Around 1930 Princeton mathematician Hassler Whitney named this concept TSP.

Considerable contributions in refining algorithms where made by George Dantzig, Selmer Johnson and the renown Delbert Fulkerson at Santa Monica's Rand Corporation in the 1960s.

Until this point advancements in computers and algorithms led to more and more cities or in our case, stops or individual addresses being considered for solution inclusion.

Dr. Arora broke ground when he found or rather envisioned a way to restate the TSP concept as a calculus equation to which the first branch of calculus (the derivative) would be useful.

Upon solving the equation he found that as friend(x) where x=optimal solution, approaches infinity (considering the many variables and with a permissible solution acquisition time of 1 hour) the result was 85%.

This meant that regardless of available computational power or algorithm elegance, the best result that can ever be achieved is 85%.

You should look deeper into this particularly in the field of cognitive research as scientist there have shown repeatedly that humans with no previous training can outperform the best computer and best algorithms, arriving at solutions 92 to 97 percent of optimal within 5 minutes.

The best computers would take a month to match that.

And computers don't provide an 85% result constantly. Remember 85% is the very best you'll ever get. Most times its between 30 to 60% of optimal. Just change 1 thing in a DOL and rerun Orion and see how many different results you get. Then consider the nearly infinite variables a driver takes into account each and every day.

So now, cognitive research is centered on discovering what enables humans to do such a good job with so many variables in such a short amount of time. If an algorithm could be formulated that would be half as good imagine the applications.

The defense industry has.

By the way, Dr. Arora's work has been authenticated by the mathematical and computer community. Hence the Godel Prize in 2008.


A bit long winded, I know. But check his research and you'll see that Orion is nothing but a blind alley.

But this can't be right! Real research be damned! According to Dragon, we are just a bunch of lazy,overpaid drivers! How did UPS get to be the International billion dollar company it is today? Oh yes, must be the great management they have in place, despite us! I forgot, why are there drivers anyway? I'm sure they could do it all without us! Who remembers the summer of '97?

 

[Dr.Brownz]

But this can't be right! Real research be damned! According to Dragon, we are just a bunch of lazy,overpaid drivers! How did UPS get to be the International billion dollar company it is today? Oh yes, must be the great management they have in place, despite us! I forgot, why are there drivers anyway? I'm sure they could do it all without us! Who remembers the summer of '97?

That was the summer my dad and I spent alot of time camping. He was a driver too

 

[SignificantOwner]

For a detailed explanation just google TSP or see Wikipeda.

I'll give you the basics. TSP (Traveling Salesman Problem) on which Orion is based originated in the 1800s as Irish mathematician William R. Hamilton considered a way to compute the shortest route a salesman could take upon leaving his hometown, calling on his accounts and returning home.

Later British mathematician Thomas Hinkman's work led to Karl Menger proposing a brute force algorithm (computing every possible route and selecting the one with the shortest distance).

Around 1930 Princeton mathematician Hassler Whitney named this concept TSP.

Considerable contributions in refining algorithms where made by George Dantzig, Selmer Johnson and the renown Delbert Fulkerson at Santa Monica's Rand Corporation in the 1960s.

Until this point advancements in computers and algorithms led to more and more cities or in our case, stops or individual addresses being considered for solution inclusion.

Dr. Arora broke ground when he found or rather envisioned a way to restate the TSP concept as a calculus equation to which the first branch of calculus (the derivative) would be useful.

Upon solving the equation he found that as friend(x) where x=optimal solution, approaches infinity (considering the many variables and with a permissible solution acquisition time of 1 hour) the result was 85%.

This meant that regardless of available computational power or algorithm elegance, the best result that can ever be achieved is 85%.

You should look deeper into this particularly in the field of cognitive research as scientist there have shown repeatedly that humans with no previous training can outperform the best computer and best algorithms, arriving at solutions 92 to 97 percent of optimal within 5 minutes.

The best computers would take a month to match that.

And computers don't provide an 85% result constantly. Remember 85% is the very best you'll ever get. Most times its between 30 to 60% of optimal. Just change 1 thing in a DOL and rerun Orion and see how many different results you get. Then consider the nearly infinite variables a driver takes into account each and every day.

So now, cognitive research is centered on discovering what enables humans to do such a good job with so many variables in such a short amount of time. If an algorithm could be formulated that would be half as good imagine the applications.

The defense industry has.

By the way, Dr. Arora's work has been authenticated by the mathematical and computer community. Hence the Godel Prize in 2008.


A bit long winded, I know. But check his research and you'll see that Orion is nothing but a blind alley.

ORION doesn't have to be perfect. It just has to be decent. Just knowing that "someone" will second guess your miles will change decision-making across the board. Gone is the old system that encouraged running up miles to increase planned time.

 

[novadriver]

I just use ODO, I dont want to be out till 930 driving in circles.. with 100 backs

 

[Harry Manback]

I just use ODO, I dont want to be out till 930 driving in circles.. with 100 backs

Surely, you meant to say, "RDO".

 

[novadriver]

Surely, you meant to say, "RDO".

yes lol

 

[soberups]

ORION doesn't have to be perfect. It just has to be decent. Just knowing that "someone" will second guess your miles will change decision-making across the board. Gone is the old system that encouraged running up miles to increase planned time.

And in its place is the new system that encourages running up wasted hours to increase a compliance metric.

 

[SignificantOwner]

And in its place is the new system that encourages running up wasted hours to increase a compliance metric.

It's the UPS way. You want us to hit a number? OK it could involve some silliness but we'll hit the number.

 

[Dr.Brownz]

Surely, you meant to say, "RDO".

We don't have PAS or EDD or anyy of that. What is RDO so I can be prepared for when the storm hits?

 

[soberups]

It's the UPS way. You want us to hit a number? OK it could involve some silliness but we'll hit the number.

I wont.

 

[clean hairy]

Which was able to calculate more correct answers when Jeopardy had the computer playing against humans?

 

[Signature Only]

ORION doesn't have to be perfect. It just has to be decent. Just knowing that "someone" will second guess your miles will change decision-making across the board. Gone is the old system that encouraged running up miles to increase planned time.

Yes, gone is the old system in which employees were treated as "partners" Now everyone is dishonest, everyone is trying to steal, everyone (at least in management) is angry about reduced benefits, added responsibilities and limited opportunities.

Yes. The system doesn't have to be perfect. But it should be humane.

 

[soberups]

It's the UPS way. You want us to hit a number? OK it could involve some silliness but we'll hit the number.

I'm not here to play silly number games. I'm here to run a delivery route safely and efficiently and to perform my duties in a manner that best represents the interests of the company, per Article 37 of the contract. If I wanted to manipulate irrelevant metrics for a living I would have gone into management.