Bayesian version of the model described in
E. Adams and M. Mesterton-Gibbons. 2003. Lanchester’s attrition models and fights among social animals. Behavioral Ecology 14:719-723.
This software was developed to support a manuscript in development (first author Elizabeth Clifton). More details about how this program can be used will be added once the paper has been submitted.
The program requires that the Boost libraries program_options and filesystem be compiled. Assuming that Boost was installed at $HOME/boost_1_73_0, the commands below can be used to compile these libraries:
cd $HOME/boost_1_73_0
./bootstrap.sh --with-toolset=gcc --with-libraries=program_options,filesystem
./b2 cxxflags="-std=c++11"
The libraries will be located in $HOME/boost_1_73_0/stage/lib. If you're using a newer version of Boost, modify the lines above and those in src/build.sh file accordingly.
Assuming you've cloned this repository into $HOME/termite-battle, build the termite battle program using the command
cd $HOME/termite-battle/src
. build.sh
If compiled and linked successfully, the program will be at src/tb.
Run the following to see command line options:
./tb --help
There is an example directory containing a example.dat file (containing the data for two battles) and battle.conf (the control file used by the program). All options listed by running the help command (see above) can be used in the battle.conf file to avoid long command line invocations.
The example.dat file provided contains data for two battles:
Battle 1: X Y
0 40 10
5 40 6
10 40 4
15 40 1
20 40 0
Battle 2: X Y
0 25 25
5 19 20
10 18 17
15 18 17
20 15 16
25 13 15
30 11 13
35 9 13
40 9 10
45 6 10
50 6 8
55 5 7
60 4 7
65 1 6
70 1 6
75 0 6
The first line of each battle must begin with the keyword Battle, followed by a battle number (arbitrary, but must be a positive whole number), followed by a colon (:), followed by the names of the two armies separated by whitespace.
The subsequent lines contain three numbers: the time, the number of individuals remaining in the first army, and the number of individuals remaining in the second army. The last line of a battle does not need to end in one army having zero individuals, but the model does not allow recruitment of new individuals, so once an army reaches zero its numbers cannot then grow again.
The example configuration file provided, battle.conf, contains some of the possible program options. Note that everything after a # character is treated as a comment:
datafile = example.dat # name of the data file to read
outfile = test # prefix used in output file names
replace = yes # OK to replace output file if it already exists?
battle = 1 # include battle 1
battle = 2 # include battle 2
saveevery = 10 # determines how often to save parameters to output file
burninevery = 1000 # determines how often to report progress during burn-in
reportevery = 1000 # determines how often to report progress during sampling
nsamples = 10000 # number of samples to save to parameters file
nburnin = 1000 # number of burn-in iterations to perform
nstones = 0 # number of steppingstones
seed = 35101 # pseudorandom number seed
fixlambda = 1.0 # fixes lambda parameter
#fixtheta = 1.0 # fixes theta parameter to specified value
#fixR = 1.0 # fixes R parameter to specified value
#fixalpha = 0.1 # fixes alpha parameter to specified value
stan = none # none, equal, or full (default is none)