Evolife

Jean-Louis Dessalles


  1. Overall description
  2. How to execute Evolife (+ download)
  3. Evolife source files
  4. Evolife classes
  5. Evolife functions
  6. How Evolife works
  7. Create your own Evolife scenario
  8. Use Evolife’s components separately (graphics, ecology, genetic algorithm)




4. Evolife Classes

In the figure below, solid lines represent inheritance links. Dotted lines indicate that a class is used in another. The schema reveals the organization of Evolife in well-separated modules.

The first module is centred on Individual. An Individual inherits its genes from Genome. Note that genes (defined in Genome) are logical segments located on the DNA. Therefore Genome inherits from DNA. Individuals have also a phenotype, defined in Phenome. An individual remembers its social links, which are stored in Alliances. Individuals gather in groups (class Group), in which they interact and reproduce. Individuals may move from group to group. Population defines a set of groups.

Storage, Base_Examiner, Examiner and Meta_Examiner are used to store numbers (simulation results) and perform statistics. Evolife_Observer is aware of the kind of data that are processed (genomes, scores, ...).

Any particular scenario should inherit from Default_Scenario (as indicated for Void_Scenario in the figure). Besides particular functions defined in the scenario, Default_Scenario provides access to simulation parameters (through inheritance from Parameters) and to the logical definition of genes on the DNA string (through inheritance from Genetic_Map).

The following figures describe the organization of the graphic modules.

The graphic module offers various windows that are used to control and visualize simulations. Some of them can be easily reused independently from Evolife (e.g. Image_Area, Draw_Area, Plot_Area, Ground). Plot_Area displays curves showing the evolution of simulation results through time (e.g. average value of genes). Genome_Area displays the genome of all agents in the population. Field_Area displays agents on a 2-D space. Network_Area displays social links between agents.
In the logic of QT, areas (QT’s scenes) require a view to be displayed. This is the role of the widget AreaView and its descendents. Active_Frame adds keyboard shortcuts, so that the same keyboard actions will be available from any Satellite_window (windows that display genomes, trajectories, field, ...).

The next figure illustrates the organization of the control windows, which constitutes the fifth and last module (the new classes are defined in Evolife_Window.py).

Evolife’s main window is Evolife_Frame. The class Evolife_Batch is an alternative to it, which can be used in batch mode when no display is wanted. Simulation_Control offers an interface with Simulation (the simulation thread, when used (not used in the 2012 version and later)). Simulation_Control_Frame offers control buttons, such as [Run] and [Step]. Simulation_Frame adds a graphic zone that displays curves. It can be reused in other programmes. Evolife_Frame inherits from it, adding concepts such as a genome window or a trajectory window, which are specific to Evolife.