MATSim Code

top using location choice scoring.

The KTI approach to duration scoring is used (desired durations per
activity types rather than per activity: this forbids using load
penalties as they are implemented in scoring, but otherwise the
"preferences" in the "preference file" --- aka desires --- do not make
sense.)

modify to allow getting params from CharyparNagelScoringParams if no desires

for MATSim2010, the person-specific parameters will be put in
person-specific CharyparNagelScoringParams.

allow to configure scoring per person without having to use location choice scoring function

not even too hackish!

looks like getting rid of ScoringFunctionAccumulator is not for tommorow

(hopefully) temporary solution: implement both the old and new
interfaces for scoring elements.

sketch for the new standard scoring function for IVT

I'm getting some problem with some elements using SumScoringFunction and
others using ScoringFunctionAccumulator: let's see whether we can get
rid of ScoringFunctionAccumulator easily before writing Wrapper classes
or other dirty hacks.

attempt to modularize location choice scoring: espilons

the idea is to separate the location-choice specific part from the ugly
copy-paste of the "core" scoring function, to allow easy replacement of
the "activity" part.

also write a f2l file

nah, don't like 28974. reverting.

don't produce duplicate agent records (sorry!!)