nojhan/sho
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

better README

Browse source 
- use gitlab's markdown
- more exercises
This commit is contained in:
Johann Dreo 2019-12-15 09:30:38 +01:00
commit 896dd3617c
1 changed files with 21 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

41
README.md
View file

@ -1,6 +1,6 @@
SHO — Stochastic Heuristics Optimization SHO — Stochastic Heuristics Optimization
######################################## ========================================
SHO is a didactic Python framework for implementing metaheuristics SHO is a didactic Python framework for implementing metaheuristics
(or evolutionary computation, or search heuristics). (or evolutionary computation, or search heuristics).
@ -12,8 +12,10 @@ The framework implements a simple sensor placement problem
and handle metaheuristics manipulating solutions represented as and handle metaheuristics manipulating solutions represented as
numerical vectors or bitstrings. numerical vectors or bitstrings.
Author: Johann Dreo <johann.dreo@thalesgroup.com> ⓒ Thales group
Executable Executable
========== ----------
The main interface is implemented in `snp.py`. The main interface is implemented in `snp.py`.
New algorithms should be integrated within this file and the interface should not be modified. New algorithms should be integrated within this file and the interface should not be modified.
@ -24,14 +26,14 @@ and a greedy search trajectory for a simple problem with only two dimensions.
Architecture Architecture
============ ------------
The design pattern of the framework is a functional approach to composition. The design pattern of the framework is a functional approach to composition.
The goal is to be able to assemble a metaheuristic, by plugging atomic The goal is to be able to assemble a metaheuristic, by plugging atomic
functions in an algorithm template. functions in an algorithm template.
Operators
--------- ### Operators
The base of the pattern is a function that contains the main loop The base of the pattern is a function that contains the main loop
of the algorithm, and call other functions called "operators". of the algorithm, and call other functions called "operators".
@ -40,8 +42,8 @@ Example of those algorithms are in the `algo` module.
For instance, the `random` algorithm depends on an objective function `func`, For instance, the `random` algorithm depends on an objective function `func`,
an initialization operator `init` and a stopping criterion operator `again`. an initialization operator `init` and a stopping criterion operator `again`.
Encoding
-------- ### Encoding
Some operator do not depend on the way solutions are encoded Some operator do not depend on the way solutions are encoded
(like the stopping criterions) and some operators do depend on the encoding. (like the stopping criterions) and some operators do depend on the encoding.
@ -49,8 +51,7 @@ The former are defined in their own modules while the later are defined
in the module corresponding to their encoding (either `num` or `bit`). in the module corresponding to their encoding (either `num` or `bit`).
Interface capture ### Interface capture
-----------------
As they are assembled in an algorithm that do not know their internal As they are assembled in an algorithm that do not know their internal
in advance, an operators needs to honor an interface. in advance, an operators needs to honor an interface.
@ -68,11 +69,11 @@ The implicit rule is to use positional arguments for mandatory parameters
on which the operator is defined, and keyword arguments for parameters on which the operator is defined, and keyword arguments for parameters
which are specific to the operator. which are specific to the operator.
Exercises
=========
Setup Exercises
----- ---------
### Setup
To setup your own solver, first copy the `snp.py` file and rename it To setup your own solver, first copy the `snp.py` file and rename it
with your name, for instance `dreo.py`. with your name, for instance `dreo.py`.
@ -87,12 +88,14 @@ The `snp.py` file shows how to assemble either a numeric greedy solver
or a bitstring greedy solver. or a bitstring greedy solver.
Implement a simulated annealing ### List of exercises
-------------------------------
Implement an evolutionary algorithm Most exercises consists in adding a single function in an existing module
----------------------------------- (or your own module) and use assemble it in the main executable.
Implement an expected run time empirical cumulative density function 1. Implement a simulated annealing.
------------------------------------------------------------------ 2. Implement an evolutionary algorithm.
3. Implement an expected run time empirical cumulative density function.
4. Implement a simple design of experiment to determine the best solver.
5. Provide a solver for a competition.