Trivia : companies that use Artificial Intelligence in bus timetables

[Simon75]

How many companies use Artificial Intelligence (AI) in bus timetables?

 

[Dai Corner]

First do, for one.

Roger Ford wrote about it recently

Bus times driven by AI

Thursday 4th January 2024 My very first work experience as a fresh faced management trainee almost fifty years ago was a placement in the schedules office at Belle Isle bus depot in Wakefield where…

busandtrainuser.com

Now, the latest development being championed by First Bus, and it seems also used by Stagecoach, is to embrace the world of ‘Artificial Intelligence’ to analyse the wealth of data now captured every hour about on the road punctuality performance to compile timetables, schedules and duties which reflect varying traffic conditions.

It was reported in the trade press just before Christmas First Bus is spending £4.5 million with a company called Prospective – “we enable fleet operators, infrastructure managers and asset owners to run their operations in sync with the city around them” or as First explains “the new system will mean that local teams can be more agile with frequent, subtle changes to ensure timetables remain accurate throughout the year”.

My local service, Stagecoach South Wales X15, has gained some random-looking variations in the late afternoon which may be AI originated.

 

[duncombec]

I suspect it will be hard to know what is generated by AI, versus what it automatically generated by scheduling software based on existing patterns.

Arriva Kent & not Surrey seem to have some interesting variations of frequency - see below - since starting to use Optibus (switching from Trapeze), Metrobus also have similar as do the latest Uno timetables on Mondays to Fridays, but I don't know what scheduling software they use.

"AI" is the new buzzword, and I suspect a fair amount of these "AI" solutions may be little more than an improvement in programming that allows it to process data it already has in a different way.

On Mondays to Fridays, my local service now leaves town at frequencies of (starting at 0613) 35-18-15-14-12-14-13-12-every 12 minutes-9-10-every 12 minutes-14-15-14-11-14-16-15-15-15-15-15-15-13-15-15-15-15-15-15-20-20-every 30 minutes to close (23:15), which of course doesn't detail any variations along the route (of which there are many, including in the early part of the frequency block and the PM peak 15 minute departure times. Really not sure why there is a sudden increase of traffic requiring a 4 minute longer journey time at 11:30 in the morning (which seems to be reason for the random 9/10 blip in the otherwise 12 minute departure times).

 

[Megafuss]

I'd be interested to know if AI can work out which trips NOT to look at when reviewing historical data, such as driver changeover times. Otherwise you will end up with some of the ludicrous headways we've seen....

I think AI can be a useful tool, but it will never replace local knowledge and many of the tools available to operators, and more importantly, properly trained staff, can generate entire functioning timetables in a couple of hours. I'd much rather have the trained specialist at a leading operator like Go South Coast or NCT creating me a timetable, than AI.

 

[Andyh82]

First using AI to write timetables was covered by BBC News last week

How AI is helping to prevent three buses turning up at once

Bus firms are now using artificial intelligence to design their timetables.

www.bbc.co.uk

Being AI and hence modern and trendy, it was given a positive write up

 

[greenline712]

Roger French, not Roger Ford! And the title is nothing trivial . . . this has the prospect of losing passengers markedly, all in the name of attempting to match timetable time to actual time.

If I go into town, intending to spend an hour shopping and then returning home on my local bus that runs every 30 minutes, I will know my return bus times. If I meet someone and delay my departure by an hour, I will still know my departure times.
Uno's 602 has MF departure times ex Watford at 1022; 1052; 1126; 1201; 1226; 1259; 1328; 1359; 1426; 1458. Timings before and after these are similarly tweaked, but this is understood for school times and peak traffic. However, running times are exactly the same (at 2hours 20 minutes through to Hatfield) from 0952 to 1548.

The only reason that Firstbus are doing this is to resolve remote monitoring difficulties by Commissioners' monitors . . . timings can be monitored by DVSA bods from a computer now; no need to actually go and look for themselves. There is no intention to "help" the passenger in this . . . none at all.
Quite why Uno have done this . . . I've no idea at all.

There is never any substitute for a schedule compiler actually checking any automagically produced timetable . . . and these prove it.

I despair . . . I really do.

 

[PG]

There is never any substitute for a schedule compiler actually checking any automagically produced timetable . . . and these prove it.

I despair . . . I really do.

One can only assume that the Head Office bean counters have seen a potential saving, over the longer term, in employment headcount by allowing non-local staff to produce timetables and rosters.

Just to clarify I think this is a daft idea!

 

[Dai Corner]

Roger French, not Roger Ford!

Oops! My apologies to both Rogers. I read their respective contributions with great interest.

 

[Baxenden Bank]

From the article quoted at #2:

“the new system will mean that local teams can be more agile with frequent, subtle changes to ensure timetables remain accurate throughout the year”.

Is that what customers want? Constantly changing timetables. How does that square with the need to register services (except for minor changes).

Basically it reads to me that they are looking to a future point when things will be totally flexible, because 'everyone' has the app and can get immediate up-to-date information. No need for a traditional fixed timetable that you could plan around. Instead you will finish whatever it is you are doing, look at an app and, hey there's a bus in five minutes I'll catch it, or, bother they brought it forward, I missed it, next bus is in 30 minutes, I'll take a taxi'. Which is fine if you can afford to instantly go with an alternative to the bus, or are prepared to wait an excessive time until the bus company deigns to send one along to mop up the remainders.

I'm not sure how this works on the ground. Many years ago PMT / Badgerline / First launched, with some fanfare 'BICCS'. This was a satellite based tracking system. It will allow us to turn buses round to get them back on time they said. Bit of a bother if you happened to be waiting farther along the route and all the buses were being turned short.

 

[Dai Corner]

From the article quoted at #2:

Is that what customers want? Constantly changing timetables. How does that square with the need to register services (except for minor changes).

Basically it reads to me that they are looking to a future point when things will be totally flexible, because 'everyone' has the app and can get immediate up-to-date information. No need for a traditional fixed timetable that you could plan around. Instead you will finish whatever it is you are doing, look at an app and, hey there's a bus in five minutes I'll catch it, or, bother they brought it forward, I missed it, next bus is in 30 minutes, I'll take a taxi'. Which is fine if you can afford to instantly go with an alternative to the bus, or are prepared to wait an excessive time until the bus company deigns to send one along to mop up the remainders.

I'm not sure how this works on the ground. Many years ago PMT / Badgerline / First launched, with some fanfare 'BICCS'. This was a satellite based tracking system. It will allow us to turn buses round to get them back on time they said. Bit of a bother if you happened to be waiting farther along the route and all the buses were being turned short.

Actually, that's what I do now. Pre-Covid there were eight buses an hour into Town so I just walked to the stop and waited. Now there are 1-3 depending on the time of day so I look online and leave home at a time to minimise my wait.

 

[Baxenden Bank]

£4.5m.
The sales team at the software company must be laughing their bits off.

I hope First Group recoups the cost of the software before their customers desert them.

Then again, perhaps it is absolutely unbelievably spiffingly brilliant on the vehicle diagrams / staff duty side of things, squeezing tuppence off here and there, interworking services left, right and centre. Nope, can't see what could possibly go wrong there, importing delays from route 1 to route 2, to route 3 several trips and hours later!

The comments on the Roger French article are a mix and I doubt they reflect the views of 'the man on the street'. One post suggests that EYMS have ceased using a similar system following customer feedback.

“After carefully listening to your feedback, we’re going back to regular departures past the hour.”

The key thing to take away is that perhaps the software looks like it works where there is a high frequency service (every 5 minutes or better), so a timetable isn't necessary and a passenger wouldn't notice a bus adjusted one or two minutes either way. Translating that to a 30 minute (or lower) frequency service is another matter.

 

[Andyh82]

There just needs to be a human or another layer of programming at the end to tidy up the results

So if the program chucks out a timetable where buses run every 19-24 minutes and take 47-53 minutes per trip, ensure it is tided up so buses in the off peak run every 20 mins and take 50 mins

 

[PG]

£4.5m.
The sales team at the software company must be laughing their bits off.

I hope First Group recoups the cost of the software before their customers desert them.

My rough, back of an envelope, calculations put the break even period at ~7 years. Long enough for those who persuaded the directors to open the piggy bank to have scarpered off to pastures new, leaving First Group to ponder over why cash flow has steadily decreased at the same time as ridership has headed south. I wonder if TfL have trademarked 'Overground' as First may regret ditching it?

 

[Statto]

Roger French is quite scathing in his latest blog on companies using AI for planning timetables

BusAndTrainUser Verify

Tuesday 23rd January 2024 Technology company Prospective and First Bus are seemingly very keen to promote and publicise the use of AI in schedules and timetable compilation. Following the article i…

busandtrainuser.com

Load of my local routes now have timetables with random variations particularly from mid afternoon onwards

 

[markymark2000]

The only reason that Firstbus are doing this is to resolve remote monitoring difficulties by Commissioners' monitors . . . timings can be monitored by DVSA bods from a computer now; no need to actually go and look for themselves. There is no intention to "help" the passenger in this . . . none at all.

The Traffic Commissioner said at the Stagecoach Exeter Public Inquiry that BODs could not yet be relied upon (source and quote below). First data definitely can't be relied upon either considering the vast amount of errors on their open data. Even once reported to the right people, they refuse to fix them as they don't see giving passengers information to plan journeys, as a priority. Potteries is still using the old Crewe Bus Station stops, the old bus station shut a year ago. Staff refuse to resolve it. South Wales only updated their information in Dec 2023 as it was still telling people to use Cardiff Central Station, a stop which had been closed since Sept 2022 as Cymru division refused to update the registration and as such bus open data.

The DVSA would have no chance checking Firsts reliability against bus open data. Not a hope in hell. Too many errors.

Decision for Stagecoach Devon Ltd t/a Stagecoach South West

www.gov.uk

8. Bus Open Data Service (“BODS”)
The “Bus Open Data Service”, or “BODS”, was created by the Bus Services Act 2017 and open data legislation. Its intention is that passengers can find online reliable information on bus services which should then drive increased patronage for operator. All bus operators who run local services in England outside London are now required to publish their timetable, fare and location data to a central system operated by the Department for Transport. DVSA has access to the data as a means of monitoring bus operator service reliability. I was concerned to note in June that the BODS data indicated that around 40% of this operator’s services were not operating at all. Neither the investigating DVSA officer nor the operator considered that figure to be anything like correct.

Traffic Examiner Eggins had visited the company again. Several potential causes of the discrepancy were identified. It appears that some timetable changes had not been uploaded to BODS. This may have arisen from a discontinuity within the wider Stagecoach group. BODS means that operators now have to provide the information to three separate organisations. I was told that the information was now better but still not right. I do not understand how the BODS data can differ from the operator’s own real-time information which both derive from the same source. If an operator with the resources and motivation of this business cannot resolve the inconsistency, it would seem that BODS is some way off being able to be relied upon as a means of monitoring bus service reliability. The second hearing was attended by senior Department for Transport officials who will, I am confident, get to the bottom of the data inconsistency.

 

[greenline712]

I stand by my comment in post #6. The Stagecoach Devon enquiry decision was 15 months ago, based on observations and investigations in June 2022, so 19 months ago.

From the quote above: " All bus operators who run local services in England outside London are now required to publish their timetable, fare and location data to a central system operated by the Department for Transport. DVSA has access to the data as a means of monitoring bus operator service reliability".

We are not privy to any comments from DfT about remote monitoring, but given that BODS has been in operation for 4 years now, it is not unreasonable that operators have been given notice to get it right, and that excuses about incorrect data will no longer be tolerated. I can see no other reason for First in Berkshire acting as they have.

If anyone can contradict me with a statement from First (or any other operator), I will happily retract my comment.

 

[Cesarcollie]

I stand by my comment in post #6. The Stagecoach Devon enquiry decision was 15 months ago, based on observations and investigations in June 2022, so 19 months ago.

From the quote above: " All bus operators who run local services in England outside London are now required to publish their timetable, fare and location data to a central system operated by the Department for Transport. DVSA has access to the data as a means of monitoring bus operator service reliability".

We are not privy to any comments from DfT about remote monitoring, but given that BODS has been in operation for 4 years now, it is not unreasonable that operators have been given notice to get it right, and that excuses about incorrect data will no longer be tolerated. I can see no other reason for First in Berkshire acting as they have.

If anyone can contradict me with a statement from First (or any other operator), I will happily retract my comment.

I’m sure you’re right. But as generally the 95% target is simply impossible to achieve, First have probably decided that this system and ‘being seen to do something’ will buy them a few more years. They can tell the BOAM officer they’ve got a great new system which will improve things, it needs to bed in and so there will be a learning curve, so in six months things might have got a bit better (tick ‘improved’ box on revisit), and so it goes on. A better option for them than a series of appearances before the TC, lots of legal costs, and probably lots of fines. And all the bad publicity that goes with it. Unfortunately the punctual guidance (and many other regulation) are driving completely the wrong behaviours. Without going too far off topic, as another example, the requirement to fit £6k-£7k of audio visual and hearing loop system to a market day bus or a 20 year old schools decker will theoretically be great for the hard of hearing or blind. In practic, it will probably just result in another service withdrawn!

 

[nw1]

I wondered why the First Solent X4/X5 had a crazy irregular timetable when I last noticed it at a bus stop in December, changed from the previous 20-min service (40-min on each).

Seems a bad idea anyway. While variations around the peak (peak extras, extended journeys due to traffic) are fine, people mostly want a regular timetable.

Bluestar don't seem to have picked up on this nonsense though.