Computer route generation

DangerDarren

Well-Known Member
#1
I've been thinking about route planning based on my experience of last years lite rally. I'm a bit of a nerd and once I've got the rally book I think I can use computer programme to generate the "Best" plan, or at least the one that delivers most points per mile. Would that count as cheating? Having recently retired I've got a bit more time on hands to consider such things. Regards all.
Darren
 

FJRPilot

Brit Butt Rallymaster RBLR1000 routemaster
Premier Member
IBA Member
#2
No, certainly not Darren. I used to use Microsoft Autoroute (2010) and just like Garmin Basecamp now it had an 'optimize' button. Just bung in any bonus locations hit the button and in theory it calculates the most mile efficient route. However like most computer programmes this doesn't take into consideration factors such as traffic, realistic road speeds etc. It does though give you an idea of what makes an efficient route which you can then manually tweak based on experience.

On previous Brit Butt rallies when they had conventionally pointed bonus locations my plan was always just to colour code each bonus location after flicking through the rally book and I always divided the bonuses into high, medium and low values depending on the points value of each one. I give them red blobs (high), blue blobs (medium) and black blobs (low). I can then see when looking at the zoomed out map where all the big value bonuses are and then plan my route around that. Obviously other factors such as restricted timings etc can all make this a bit tricky too. :rolleyes:
 
#4
Mark is right lots of mapping / computer programmes that are design to help. I group my controls by value as well pretty sure most people do. Once you've done this its good to get it on mapping programme and get an overall picture.
then comes the master computer in your head, really this is the only one now that matters and it makes a decision on which route , I say to myself that looks a good area direction then plan from there. I never press the optimize button mainly because I don't think it will consider road conditions time of day congestion etc ,

But the big moment in any plan is that decision you make. I personally don't believe any computer will do that better than yourself. There are whole books written on how we make them decisions and the various traps that are inherent within that process I've read a few of them.
The mapping programme will let you do the detail of the route and help you decide what is feasible for yourself to achieve
 

Stephen!

Flivver Flyer
Premier Member
IBA Member
IBR Finisher
#5
No, certainly not Darren. I used to use Microsoft Autoroute (2010) and just like Garmin Basecamp now it had an 'optimize' button.
Have you really had any good luck using "Optimize" in BaseCamp? I've seen it take perfectly crafted routes and totally mangle them into something two or three times as long as the original. In fact, I've never seen a successful BaseCamp optimize...
 

IBRX3

Premier Member
IBR Finisher
#6
Don't want to dampen your enthusiasm, but if you get basecamp to take into account bonus time restrictions and individual points whilst optimising that's a new one.
There is a work around in basecamp to get it to automatically sort bonuses from high points to low points in a list so you can give each category a symbol , but after that you have to eyeball it. Optimise is no use in most cases as arriving at one big bonus at the wrong time can screw the whole route
 

Robert

IBR Finisher
Premier Member
#7
The underlying routing problem is the Travelling Salesman Problem (TSP). Its mathematical solution is one of the most complex challenges in mathematics and in logistics. I had considered delving deeper into the subject but it got clear very soon that this is a massive task and would require a sound background in handling algorithms, programming, linking them to open source maps and interfacing....
I concluded that for our particular problem of rally routing, the human supercomputer aka brain is still the best. Just learn how to introduce the data in an efficient way into basecamp, develop promising routes and use your experience and knowledge as a filter to select the right one.
 

Robert

IBR Finisher
Premier Member
#11
I get your point. Indeed, this is an asset. There are other sites like this one to optimize the order of selected points, however the fundamental challenge will likely always be for humans: to select a number of points with certain values from a pool of possible locations in a given timespan. There, the experienced rider comes in. And I think we should be happy that this task can't be replaced by a software ;)
 
#12
Well...it *can* route to a set (max: 20 for free) of random locations, creating an "ideal" route between them.

Paste them in the 'import' window, then execute.

http://www.routexl.com/f=754R5gFEhN1c

Fifteen counties in my state; here's routing to all of the county courthouses.
yes it has optimised them but a fairly simple route. once you throw in complications such as time restrictions and more controls than can possible be done all of vary values and combinations of controls that can be done for bonus's then it sort of isn't going to cope, and if it did the rally master would soon throw a spanner in the works. ultimately it wont decide for you what controls not to go to.
excel can be used to sort and organise your controls in to some sort of value order fairly simply
but for me I leave the optimise button well alone, in fact I don't always like what the route the planners picked so go different routes sometimes it even works, when it comes to your decisions and ultimately that's what it comes down to, if you have a feeling its not a good idea it probably isn't.
 

kwthom

=o&o>
Premier Member
IBA Member
#14
No no no...please don't misconstrue my initial comments on computer-generated routing. The capability to do this exists, yes, but @Robert comment is dead-on. The computer can't take into account any of the outside variables that we as riders have to manage.
 
#15
Computers, particularly when used in near-realtime, such as WAZE and Garmin's live traffic via SmartPhone Link, are useful for planning routes while on the road. In urban areas where you lack local knowledge, these tools are powerful but you need a reasonability test in your mind. All algorithms are subject to failure. Most optimisers produce a good result or an optimal result. There is a reason Google limits you to 10 locations on its routing. RouteXL does well will 100 locations but it is a good result, optimisation with that number would take until the end of time.

The possible routes in the classic travelling salesman problem expands as the factorial of the number of locations. Specifically (n−1)! / 2. For say 11 cities the combinations are almost 20,000,000. The factorial of 100 is
93, 326, 215, 443, 944, 152, 681, 699, 238, 856, 266, 700, 490, 715, 968, 264, 381, 621, 468, 592, 963, 895, 217, 599, 993, 229, 915, 608, 941, 463, 976, 156, 518, 286, 253, 697, 920, 827, 223, 758, 251, 185, 210, 916, 864, 000, 000, 000, 000, 000, 000, 000, 000

That is a BIG number but you only have half that many.
 
Last edited:

GSears

Dambuster... Bouncing panniers...
Premier Member
IBA Member
#17
Having ridden along 2 different roads that my 590 assured me were closed (via smartphone link) on the way to the first UK RTE of the year, I've now disconnected the smartphone option. It wasn't that smart o_O
 
#18
Having ridden along 2 different roads that my 590 assured me were closed (via smartphone link) on the way to the first UK RTE of the year, I've now disconnected the smartphone option. It wasn't that smart o_O
It works pretty well in the southwestern U.S. but indicating a road is closed is a common problem. Twice I have been riding north on the 15 from L.A. to Las Vegas and been told I-15 was closed. There is no f'ing alternate and there were not 45000 cars backed up so I soldiered on...

On the otherhand, on I-70 in Utah it saved me hours of delay behind a massive accident.
 

Robert

IBR Finisher
Premier Member
#19
In terms of satellite navigation, America and other locations are poles apart. Garmins were developed for US roads, and they usually do funny things when they are confronted with the European road system, especially when dealing with small roads.

Routing in the US is damn easy :). I don't know how many times I was shouting at my sat navs here in Europe.....:rolleyes::D

Finding a good route for given number of points is quite simple (see the example above:p). Our problem is even more complex: WHICH points to select for a given time? This is the mathematical overkill o_O
 
#20
In terms of satellite navigation, America and other locations are poles apart. Garmins were developed for US roads, and they usually do funny things when they are confronted with the European road system, especially when dealing with small roads.

Routing in the US is damn easy :). I don't know how many times I was shouting at my sat navs here in Europe.....:rolleyes::D

Finding a good route for given number of points is quite simple (see the example above:p). Our problem is even more complex: WHICH points to select for a given time? This is the mathematical overkill o_O
DAMN you are a rally monster in Europe. I expect your GPS experience is unequalled. When I lived in Spain (Costa Brava, just north of Barcelona) my cheap little Garmin was a godsend. Was it always right - NO! It did help me find things in a city (Barcelona) that was 24x7 chaos. In Paris it was awful, I often just found parking outside of town and took a cab. It was fun listening to it pronounce ES/FR street names.

I really want to ride a rally in Europe someday - SOON.