Skip to main content

Here is a link to a story about a new concept for Rail Intermodal Transport. Amazon wants to use container trains as sort of mobile fulfillment centers. The whole thing sounds sort of science fiction like, but I will reserve judgment on how well it works until it is rolled out. 

Amazon Train Idea...

Last edited by Rich Melvin
Original Post

Replies sorted oldest to newest

What we have might be called "weak artificial intelligence" because it  operates ONLY within a set of rules (machines able to play chess or assembly-line 'bots able to "learn" how to make "better" welds) whereas both the Amazon scenario and self-driving carz require what might be called "general intelligence" (the human abilities of wit, humor, morality, ethics, free will, self-awareness, intuition, and consciousness; in short, the ability to make correct decisions with insufficient data). This because both would have to operate while interacting with actual human intelligence and the only thing that can do that is.....another human intelligence.

Based on what I see with their delivery vehicles in my area; Amazon doesn’t have the first darn clue about true supply chain.  Generally speaking; supply chain practitioners try to operate in the most efficient and cost effective manner possible.  I’d hold UPS up as the model for supply chain especially in creation and management of their pickup and delivery routes.

Amazon, on the other hand, appears to have a delivery model that is the antithesis of efficiency. There are days I’ll see as many as half a dozen different Amazon delivery vehicles in our neighborhood throughout the day.  I think about the only thing saving Amazon’s screwed up delivery process at this point is their use of contractors and sub-contractors.

It appears to me that Amazon is willing to absorb higher distribution costs right now in anticipation of drone and self-driving vehicle deliveries down the road.  And from my perspective after a career in industrial logistics and supply chain; neither will be the panacea some believe they will be.

Curt

Last edited by juniata guy
juniata guy posted:

...it strikes me as millennial nonsense...

AMEN.

mlaughlinnyc posted:

Whoever wrote that obviously knows nothing about operating on a railroad in the real world....

Another AMEN.

The railroad industry simply cannot (or maybe will not?) operate with the level of precision and reliability demanded by this new service idea. Not even with PSR, which really stands for "Pretty Sad Railroading."

juniata guy posted:

...Amazon doesn’t have the first darn clue about true supply chain...Amazon...appears to have a delivery model that is the antithesis of efficiency...

And a third AMEN.

I ordered a $10 product from Amazon on a Saturday afternoon and the United States Postal Service delivered it to me the next day on Sunday. The USPS employee who delivered it was being paid overtime to work on a Sunday. There is NO WAY Amazon made any money on that product.

I like the fast deliveries, but this has got to be a massive loss leader for Amazon.

Rich Melvin posted:

I ordered a $10 product from Amazon on a Saturday afternoon and the United States Postal Service delivered it to me the next day on Sunday. The USPS employee who delivered it was being paid overtime to work on a Sunday. There is NO WAY Amazon made any money on that product.

I like the fast deliveries, but this has got to be a massive loss leader for Amazon.

99% of the time, I've never paid for shipping from Amazon, nor do I pay for Amazon Prime.  I click on the option where shipping is free if I'm willing to wait 5-7 days.  The orders usually come in 5 days or less, one even came overnight!  Nothing I've ever ordered was so vital that I needed it overnight or in 1-2 days.

Rusty

Hw about bringing some passenger experts who run the Swiss railroads.  But instead of running passenger trains, they would be running high quality fast short trains for Amazon, FedEx, USPS, and UPS.

Seems the above need the kind of service the REA and RPO's used to provide.

Amtrak also should let third parties like the above run on the NEC at night.  

I could see the proposed Texas Central high speed rail run express.  I think the Frence use older TGV trainsets for express mail and express.

I doesn't take outside experts to run short fast trains.  We know how to do it, but the economics don't work.  More short trains gets very expensive fast.  Not only crew and locomotive expense.  Line capacity costs are a deal breaker in this case.  The entire American railroad network is designed for long trains with blocks of cars between classification points running daily or twice a day, except if you are the UP which runs a three track freight railroad.

European networks are designed for the kind of service they run.

The UPS, FedEx and USPS networks are designed around moving containers/trucks between major sorting centers.  None of those centers have rail sidings, and most are not near a railroad.  Many of those boxes between centers have their long haul by rail, and there are many daily trains dedicated to UPS.  For example, CSX has a daily UPS train between Worcester and Chicago.  UPS truck haul containers  to Worcester from many sorting centers in the Boston area and other points on the east coast from Rhode Island to Maine.  That is a very efficient system.

Amazon isn't worth mentioning in this context.  They have an ad hoc system cobbled together with a huge number of subcontractors.  You can always get overnight delivery if you throw enough money at it.

 

A business which has no customers has value to no one, especially the stockholders.  Seems the railroads are operating in a way to "drive" customers off.

Also, many have said one of the large costs of running a railroad is the people they employ.  But what about the ROW.  The railroad has to maintain and pay taxes on it.  That is a fixed cost if you run one or many trains over it.

Lastly, short vs longer trains.  To me, it would take less time to build and tear down shorter trains in a yard than larger.  Yards would turn over faster.  And there are some yards or terminals which have difficulty with moving traffic.  Houston is one.  Seems there are always trains parked somewhere on the line along the Hardy Toll Road.  It has two main tracks.  Recently, UP added a siding to help make the line more fluid.

Last edited by Dominic Mazoch

Like many ideas this proposed idea from Amazon is simply that, an idea. Look through the backlog of ideas or inventions that have been patented and you see a lot of 'great ideas' that frankly are ludicrous if you look at them now...and what they propose is not going to happen in the short term, I can guarantee that. 

However, it is doable, but requires some fundamental things that don't exist and may not be able to be overcome:

1)Given Amazon's model, that rail would even work for them. With the 1 day delivery or intra day delivery they are banking on having the inventory local, they have local fulfillment centers that allow rapid delivery using the current model (though  I have had 1 day delivery on items that were shipped from across the country). The problem with rail, especially in this day and age of "precision railroading" or whatever they call it, is that rail shipping if I read right in trains magazine, has gotten even less time conscious. 

Does Amazon think that they are going to have a train full of shipping containers full of various items (sort of like their current distribution centers) and as people order, they check where the nearest of these trains is, and ship from there (the description sounds like that)? If so, then Amazon's plan is kind of like the 1950s idea of having trains with missiles on them continually running the rails,  in their case it will be having trains with Amazon containers running a kind of continuous route  and having trains all over the country doing this running some sort of regional loop(kind of like running around one of our layouts).....and I wonder if they plan to restock the containers on the fly, too...

Just don't know if rail could support this,  if traffic would allow it or if the railroads could handle having trains running like that. 

2.The future of drones is cloudy as well, given how much Amazon delivers on a given day, just not sure having so many drones working in local airspace would work. Too many chances of accidents, Drones being blown into power lines, Drones dropping packages from a height and killing someone because of a malfunction, and to be honest, swarms of drones in neighborhood delivering packages sounds very intrusive. Sure, you get deliveries in neighborhoods all the time via UPS, Fedex, the USPS, Amazon, but it isn't the same thing of having airspace full of drones, very different level of annoyance and yes, risk. 

3)Can this be run by AI? The self repair by robots is still questionable, there has been limited success with robotic and self repair, but it isn't quite science fiction, it is just really early. Likewise, a distribution system like this will require sophisticated tracking and dispatching capabilities to make it work, and as good as AI/machine learning systems can be in this area, it isn't there yet....but could be, and not that far in the future. 

AI systems don't need to have emotions to function, they don't need wit or humor to be able to solve problems, they need sophisticated self learning systems to learn as they go the way people do, things like intuition and human's ability to put together a bunch of guesses and inference to solve a problem are not easily replicated by AI, but they don't have to be, AI intelligence is much like work on animal intelligence, what they have found is that  animal intelligence uses very different areas of their brain working in different ways to do things humans do very differently; AI wouldn't replicate human processes, it would find a different way to get the same results, and has.

Put it this way, AI is used all over the place in ways people don't even know, it is used in everything from appliances and a lot of the trading going on in financial markets have AI components to them, including machine learning, it is used to solve problems that humans haven't been able to because it works differently *shrug*. Rule based AI has been around a long time, what is being developed right now is not rule based, it is based in self learning and in mimicking how humans solve problems, and it has advanced a lot in the last 5 years or so from what I can tell. 

Does this mean Amazon's idea will work? If it does, it has some fundamental problems that would definitely block it today, and not all of it is technical in nature, there is a lot of human engineering that would allow this to work, including the railroads, not exactly paradigms these days of innovative thinking or thinking outside the box, and in people accepting this kind of system. I agree that Amazon's current model likely is not going to work long term with shipping, it works now because Amazon makes money a lot more from its cloud businesses than the retail side, but as others have side, they are expecting it, like their warehouses, to become more and more automated, to really make money with their model of almost on demand shipping. 

 

Aaaahhh  no , will never happen 

Sorry , gotta be some Buck Roger's idea from someone who really doesn't understand how the real world  of railroad logistics works.

I always thought we would go back to a Railway Express type container with sliding side doors to be picked up and dropped off daily. But with the railroads eying mega-trains more and more ,these special delivery type dispatches won't happen either.

In my opinion that will be the railroads only way to compete with  truck express shipments. But then again that would require more people and individual handlings like they want to separate themselves from. 

Railroads and small package delivery are two unrelated kinds of business.  In the mid 20th century, railroad LCL and Railway Express lost huge amounts of money that would have been better spent improving what railroads do best.  The railroad business is about moving large lots of goods (100 ton cars) in large batches (trains) long distances very efficiently.  That includes moving large numbers of containers in dedicated trains, UPS trains for example.  The Amazon thing simply does not fit the profitable railroad business model.
 

Add Reply

Post

OGR Publishing, Inc., 1310 Eastside Centre Ct, Suite 6, Mountain Home, AR 72653
800-980-OGRR (6477)
www.ogaugerr.com

×
×
×
×
Link copied to your clipboard.
×
×