Context Navigation

#177 closed task (done)

Restructure genetic programming plugins

Reported by:	gkronber	Owned by:	gkronber
Priority:	low	Milestone:	HeuristicLab 3.3.0
Component:	Functions	Version:	3.2
Keywords:		Cc:

Description (last modified by gkronber)

Create a new plugin structure for the genetic programming parts.

Change History (15)

comment:1 Changed 16 years ago by gkronber

Milestone changed from 3.1 to 3.2
Version changed from 3.1 to 3.2

comment:2 Changed 16 years ago by gkronber

Status changed from new to assigned

comment:3 Changed 16 years ago by gkronber

Description modified (diff)
Summary changed from Restructure functions plugin to Restructure genetic programming plugins

plugins involved:

HeuristicLab.StructureIdentification
HeuristicLab.Functions

comment:4 Changed 16 years ago by gkronber

created a branch for the refactoring with r643.

comment:5 Changed 16 years ago by gkronber

created new plugins GP, GP.StructureIdentification, GP.StructureIdentification.Classification and GP.StructureIdentification.TimeSeries in the refactoring branch and moved existing classes from HeuristicLab.Functions and HeuristicLab.StructureIdentification into the new plugins with r643:647.

New plugin layout:

HeuristicLab.GP: Contains everything that is needed to work with the (abstract syntax) tree representation of GP. This includes interfaces IFunction and IFunctionTree and base classes. Also included are manipulation and crossover operators that work on IFunctionTrees and operators that create random trees. The plugin also defines a subtype of OperatorLibrary that is useful in the context of GP.
HeuristicLab.GP.StructureIdentification: Defines functions that can be used in a structure identification modeling scenario and evaluators that can interpret the functions. Also included are exporters that can export trees to other representations (HL2, Scheme, Excel: see #306). This plugin also defines an operator that can be used to import datasets exported from HL2.
HeuristicLab.GP.StructureIdentification.Classification: Contains evaluators for classification problems and an operator that injects a default function library for classification problems (no time offsets, no differential terminal symbol)
HeuristicLab.GP.StructureIdentification.TimeSeries: Contains evaluators that are only useful in a time-series modeling scenario. Also includes an operator that injects a default function library for time-series modeling.

related (see #299)

HeuristicLab.GP.SantaFe: Implements the santa fe ant trail benchmark problem. It demonstrates how a new 'tree programming language' (which is not based on the HeuristicLab.GP.StructureIdentification plugin) can be implemented based on the classes in HeuristicLab.GP.