Commit e6489f9b authored by Gauthier Quesnel's avatar Gauthier Quesnel
Browse files

README: update with new parameters

parent 178ff3e2
Pipeline #6300 passed with stage
in 3 minutes and 8 seconds
= Baryonyx
Gauthier Quesnel <gauthier.quesnel@inra.fr>
v0.5.x, 2019-??-??
v0.5.x, 2020-01-??
:toc:
:homepage: https://github.com/quesnel/baryonyx/
......@@ -52,9 +52,9 @@ cd baryonyx
git submodule update --init --recursive
....
Default, Baryonyx provides a shared library `libbaryonyx-0.4.so` (with
hidden symbol), a static library `libbaryonyx-0.4.a` (all symbols are
public) and an executable `baryonyx-0.4`. To compile and install in
Default, Baryonyx provides a shared library `libbaryonyx-0.5.so` (with
hidden symbol), a static library `libbaryonyx-0.5.a` (all symbols are
public) and an executable `baryonyx-0.5`. To compile and install in
the default CMake install directory:
....
......@@ -146,7 +146,7 @@ make install
To run Baryonyx in solver mode (i.e. trying to valid all constraints):
....
baryonyx-0.4 file.lp
baryonyx-0.5 file.lp
....
To run baryonyx into the heuristic model (i.e. trying to valid all
......@@ -154,7 +154,7 @@ constraints and optimize the solution), add a `-o` or `--optimize`
option to the command line:
....
baryonyx-0.4 -o file.lp
baryonyx-0.5 -o file.lp
....
To run baryonyx into the heuristic model (i.e. trying to valid all
......@@ -162,7 +162,7 @@ constraints and optimize the solution), add a `-o` or `--optimize`
option to the command line:
....
baryonyx-0.4 -o file.lp
baryonyx-0.5 -o file.lp
....
The Baryonyx solver have many parameters. Some parameters are global,
......@@ -349,7 +349,7 @@ To assign parameters to solver or optimizer algorithms, use the `-p
| integer
| show information if greater than 0
| init-policy
| init-policy (solver only)
| string
| Change the initialization and reinitialization policy of the solution vector. Values are:
......@@ -361,18 +361,54 @@ To assign parameters to solver or optimizer algorithms, use the `-p
* `optimistic-solve`: found a solution for each (or or `init-policy-random`
rate) constraints. For soft constraints, affect one to the maximum variables
that valid the constraint.
* `cycle` (default): Only for the optimization mode, start with the
`pessimistic-solve` mode and change to `bastert` then `pessimistic-solve`
mode. The change between mode is done if and only if three times,
optimization fails to improve the current best solution.
| init-policy-random
| init-policy-random (solver only)
| real
| [0-1] (default, 0.5) parameter of the bernoulli's law to be used in conjunction with the `init-policy` parameter. If the law returns 1, it uses the `init-policy` algorithm to initialize `X_i`, 0 means use a toss up to choose 0 or 1 according to the `init-random` value.
| [0-1] (default, 0.5) parameter of the bernoulli's law to be used in
conjunction with the `init-policy` parameter. If the law returns 1, it
uses the `init-policy` algorithm to initialize `X_i`, 0 means use a toss
up to choose 0 or 1 according to the `init-random` value.
| init-random
| init-crossover-bastert-insertion (optimizer only)
| real
| [0-1] (default, 0.5) parameter _p_ of the bernoulli's law used when 0 occurs with the `init-policy-random`.
| [0-1] Probability to insert a bastert solution during the crossover
operation.
| init-crossover-solution-selection-mean (optimizer only)
| real
| [0-1] Probability to select a solution to do the crossover operation. This
parameter allows the selection of solution in the population. 0 means best
solution, 1 means the worst in mean.
| init-crossover-solution-selection-stddev (optimizer only)
| real
| [0-1] Probability to select a solution to do the crossover operation. This
parameter allows the selection of solution in the population. The standard
deviation for the normal probability law.
| init-mutation-variable-mean (optimizer only)
| real
| [0-1] Probability to mutate the solution after the crossover operation. This
parameter defines the number of variables to change. The mean for the normal
probability law.
| init-mutation-variable-stddev (optimizer only)
| real
| [0-1] Probability to mutate the solution after the crossover operation. This
parameter defines the number of variables to change. The standard deviatation
for the normal probability law.
| init-mutation-value-mean (optimizer only)
| real
| [0-1] Probability to mutate the solution after the crossover operation. This
parameter defines the value of the variable. The mean for the normal
probability law.
| init-mutation-value-stddev (optimizer only)
| real
| [0-1] Probability to mutate the solution after the crossover operation. This
parameter defines the value of the variable. The standard deviation for the
normal probability law.
| storage-type
| string
......@@ -400,7 +436,7 @@ files. All Baryonyx parameters are available to perform the benchmark.
For example:
....
baryonyx --bench bx-0.4 -pdelta:0.01 -ptime-limit:60 spp.csv
baryonyx --bench bx-0.5 -pdelta:0.01 -ptime-limit:60 spp.csv
....
The benchmark mode updates the `csv` file with results of computation. The
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment