Tue II: It’s progress Jim, but not as we know it

Following this morning’s efforts, redoubled the attack on Excel. And it didn’t really work.

It is now apparent that trying to simulate a TCA in Excel only is beyond* me.But the exercise was not a wasted effort, as I have still learnt a small bit about 1d TCAs.

However,the issue is now where to go next. Options include:

• coding from scratch in Python, e.g. a la pygame
• coding from scratch in VBA + Excel: this is tempting, as the lattice environment is already to go
• trying out Revolver One from  http://www.resolversystems.com/ which is a spreadsheet but based on Python, so can use Python to control what goes on with the simulation (there is a Game of Life for this already:  http://www.resolverhacks.net/game_of_life.html, which is tempting) (also see http://www.dalkescientific.com/writings/diary/archive/2007/10/18/resolver.html for a review from a Chemistry angle) (this is also tempting: http://www.resolversystems.com/sendus/)
• Visual Bots anyone?: http://www.visualbots.com/ I remember this from donkey’s ago, but never went anywhere with it …. is its time now?

Next steps will be to work out what I’m actually looking to do, then pick the tool ….

* currently

Tue I: My heads hurts …

Oh the joys of implementing something straight forward in a tool that doesn’t easily resolve to the desired design!

Trying to get the TCA model up and running in Excel: still failing.

Will leave for now and attempt again later today. I think a re-read of the MSDN blog on GoL will be very sensible (http://blogs.msdn.com/excel/archive/2007/11/02/iteration-conway-s-game-of-life.aspx)

Mon: Developing an Excel TCA …

… and it is going slowly …

Using the rules found on this page in Excel (http://www.thp.uni-koeln.de/~as/Mypage/traffic.html), but without VBA currently. I can’t see an easy way of producing rule #2 distance without VBA currently, but hoping that sleeping on it + inspiraton will produce something of use (http://www.cpearson.com/excel/ArrayFormulas.aspx) …. or offset, index, match maybe? (http://support.microsoft.com/kb/152407)

UPDATE: will a combination of min, column with a constant look ahead range dependent on Vmax work?

Some other links of possible use:

• http://www.thinkbottomup.com.au/site/blog/Game_of_Life_in_XAML_WPF_using_embedded_Python Game of life in Python with XAML / WPF
• a freeware Game of Life (GoL) simulator http://psoup.math.wisc.edu/Life32.html
• Plife GoL package in Python http://plife.sourceforge.net/
• Misc traffic info http://trafficforum.org/

Off to cook the Missus a steak sarnie for dinner !

Sat: mucking about with Excel + car info ….

Trying out the multi-agent simulation using my new found favourite website (Multiagent System via http://people.revoledu.com/kardi/tutorial/index.html) in Excel …. still ruminating on the results and how to take this forwards…..

In the meantime, some info:

• Mondeo car dimensions (http://www2.fordconnection.com/fordconnection/multimedia/gbr_en/00q/yyb/gbr_en00qyybedimspc.htm)
  • length: 4844mm
  • Overall width with mirrors: 2078mm
• Motorway width info (http://www.motorwayarchive.ihtservices.co.uk/m621leeds.htm)
  • M621 motorway width J3 to J7:  Generally the motorway, which has a design speed of 50mph, is 110ft. (34m) wide comprising dual three lane 36ft. wide carriageways and 1ft. wide marginal strips, separated by a 10ft. wide central reserve with safety fencing and flanked by two 8ft. wide hard shoulders and 4ft wide verges

Wed: Excel iteration …

Trying to follow the Excel CA tutorial I found on MSDN the other day … but haven’t quite cracked it. Tomorrow!

Tue II: more reading …

Stumbled across a page with a number of interesting tutorials on it http://people.revoledu.com/kardi/tutorial/Excel/index.html including things like solving ODE’s, kernel regression, etc., mostly from within Excel  (from the looks of things).  Interestingly the author for his thesis looked into pedestrian simulations, and does mention CAs (http://people.revoledu.com/kardi/publication/Dissertation.pdf).

Tried to get a number of Python-based CA’s working, but irritatingly can’t get the PIL imaging library to work with v2.6 of Python on Win XP, due to MS C compiler used to compile either Python or PIL …. great.  Managed to find a number of CA’s though.

NB Saw Paramics mentinoed somewhere before — they make microsimulation software: http://www.paramics-online.com/.

PS Go look at this chaps website as it could have some things of use on it ! http://www.maerivoet.org/index.php?page=phdresearch-dissertation

NB2 Not sure if this will have some useful Excel-like links in it: http://wofford-ecs.org/IntroComputationalScience/preface.htm

What I need to find out is: what is the difference between Mathematica and Matlab?

Tue I: Reading about TCAs …

Found a link that points to TRANSIMS documents that actually works: http://tmip.fhwa.dot.gov/resources/clearinghouse/docs/transims_fundamentals/. Chapter 6 for the MicroSimulator is quite good, as it explains the algorithms that are used in lane changing, etc.

This paper http://psroc.phys.ntu.edu.tw/cjp/v41/671.pdf describes a 3-lane CA model (Numerical simulations of a three-lane traffic model using cellular automata, AK Daoudia, N Moussa – Chinese Journal of Physics, 2003)

This paper ftp://ftp.esat.kuleuven.ac.be/pub/sista/bellemans/papers/01-89.pdf (Models for traffic control, T Bellemans, B De Schutter, B De Moor – Journal A, 2002) has some good definitions on:

• deterministic vs stochastic models
• discrete vs continuous traffic simulations
• differences between microscopic and macroscopic models (and refers reader to Hoogendoorn and Bovy [2000], “State-of-the-art of vehicular traffic flow modelling.” Tech. rep., Transportation and Traffic Engineering Section, Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft, The Netherlands. Accepted for publication in the Special Issue on Road Traffic Modelling and Control of the Journal of Systems and Control Engineering – Proceedings of the Institution of Mechanical Engineers, Part I.) for more detail.
• it also describes continuous, macroscopic models — Payne and Lighthill-Whitham-Richards
• some names of people — it would be good to track down their websites. Code / simulations could be posted online:
  • K Nagel
  • DE Wolf
  • P Simon
  • M Schreckenberg

This search at ScienceDirect has indicated some useful looking papers: ["three lane" traffic cellular automata] http://www.sciencedirect.com/science?_ob=ArticleListURL&_method=list&_ArticleListID=1045787687&view=c&_acct=C000063589&_version=1&_urlVersion=0&_userid=6805906&md5=4f6e8b3f9b7df599d1baeb47f3c4b453

Mon: ill = bah! …

Not feeling that tip-top today, so will be brief:

• this is a dupe, but is the original TCA paper by Nagel et al: http://hal.archives-ouvertes.fr/docs/00/24/66/97/PDF/ajp-jp1v2p2221.pdf
• Two lane simulations using cellular automata: http://arxiv.org/PS_cache/cond-mat/pdf/9512/9512119v1.pdf
• Two lane traffic rules for cellular automata: a systematic approach: http://arxiv.org/PS_cache/cond-mat/pdf/9712/9712196v1.pdf
• This could prove to be an interseting Google Scholar search: http://scholar.google.co.uk/scholar?hl=en&q=%22two+lane%22+traffic+cellular+automata&btnG=Search

I think it would be good to have a trawl through and see if any of the above have some code to pick over on a personal web site ….

Thu: tracking TCAs…

No movement on Wednesday, due to 17miles of tailbacks on the M25 + shoddy weather + late running meetings ….

… reading was performed, and these seem to warrant more investigation:

• AINSUN2 http://www.aimsun.com/hughes.pdf
• TRANSIMS http://transims-opensource.net/index.php?option=com_content&view=article&id=54&Itemid=91
• the technical report detailed here: http://www.maerivoet.org/website/phdresearch/publications/resources/doatcamfht.pdf

Tue: more CAs ….

Been seeing what is possible in R. Found Tinn-R http://www.sciviews.org/Tinn-R/ which is an editor for R (installed it, not keen, will probably refer back to Eclipse shortly …)

Been trying out some of the code in this paper http://cran.r-project.org/doc/Rnews/Rnews_2003-3.pdf that in turn relates to this paper: http://www.jstatsoft.org/v22/i09 (no ideas why WordPress / Firefox will HTML-ise one link but not the next)

And in Excel: http://www.economicsnetwork.ac.uk/cheer/ch17/hand.htm, http://dobbscodetalk.com/index.php?option=com_myblog&show=Gliders-Hasslers-and-the-Toadsucker-writing-and-documenting-a-structured-spreadsheet-Life-kit-for.html&Itemid=29 and http://www.nickpisca.com/BLAST/index.php?title=Cellular_Automata and even on the Excel MSDN blog: http://blogs.msdn.com/excel/archive/2007/11/02/iteration-conway-s-game-of-life.aspx (from Modelling in Excel www.cs.usyd.edu.au/~info1903/lab06/notes.pdf)

Seems all well and good …. now need to work out how models are represented, as results from traffic simulations appear to be presented as diagonal lines. Will this work for 2D traffic models? How do TCAs work?

http://www.google.com/url?sa=t&source=web&ct=res&cd=20&url=http%3A%2F%2Fwww.cs.usyd.edu.au%2F~info1903%2Flab06%2Fnotes.pdf&ei=kHzLSsm2Aord-QafgLXpBA&rct=j&q=cellular+automata+excel&usg=AFQjCNEc7ACnd4U0Tpk7u73RTo-8W7imZQ