TITLE
============
Analyse Genetic Programming Runs

OBJECTIVE:
To have a set of easy to be used scripts to produce commonly 
postscript graphs for the RMIT GP package.


The project will involve four phrases. 

PHRASE ONE
============
Develop a basic script to capture the following information
  - LOG FILES : run-logs
  - GRAPHS    :
    > BestFitness    vs. Generation     average method: static, all
    > AverageFitness vs. Generation     average method: static, all
    > AverageDepth   vs. Generation     average method: non-static
    > AverageSize    vs. Generation     average method: non-static
    > WorstFitness   vs. Generation     average method: non-static

PHRASE TWO
============
Design a set of scripts to re-do the experiments presented in the 
workshop paper.

. LOG FILES : gen*

. SCRIPT ONE 
  - GRAPHS    :
    > number of occurrences vs. size
    > number of occurrences vs. depth
    > number of occurrences vs. fitness
  
. SCRIPT TWO
  - GRAPH     : 
    > number of tree shapes vs. generation
    
. SCRIPT THREE
  - GRAPHS
    > number of evaluations vs fitness    and size
    > number of evaluations vs generation and size

. SCRIPT FOUR
  - STATISTIC GATHERING
    A script which analyzes the gen* files and collects information to 
    store into a *.stat file.

    Information needs is:
    # TotalEvaluation 
    # NoOfStringDistinct
    # Top5MostFrequentlyEvaluatedPrograms, Fitness, NoOfEvaluations, FirstOccuredGenNo

    # NoOfCommutative
    # MostFrequentlyEvaluationNo
    # Top5MostFrequentlyEvaluatedPrograms, Fitness, NoOfEvaluations, FirstOccuredGenNo

    # Top5BestPrograms, Fitness, NoOfEvaluations, FirstOccuredGenNo


 PHRASE THREE
 ==============
 Utilizing the newly created scripts to examine runs with decimation for 
 the problems of Santa Fe Ant, Symbolic regression, MAX and Parity and have
 a comparison with the standard approaches.


 PHRASE FOUR
 ==============
 
 - Refine the script in phrase one to add one and two standard
   deviation information 

 - Refine the script in phrase one to add multiple runs

 - Refine the script in phrase one to capture information in multiple
   directories.

 - Design a script to measure diversity of the programs 
     genotype   - program differences
     phenotype  - fitness differences

 - Find a way to convert shape into a number to analyse
   . fitness changes for the most frequently occurred shape
   . fitness changes for the shape which has the best fitness
   . plot fitness landscape with shape information

 - Find out most frequently occurred longest patterns in the 
   programs. 


. Presentation

  - What are the objectives of your summer project?
  - What have you done to achieve these objectives?
    What methodologies did you use? 
    .  Methods 
    .  Brief description of each script
    .  Validation -  experimental data on different
                     problems
 
    What are the difficulties encountered? How you
    overcome them ?
    How did you overcome them?
    What are the limitations of the stuff you developed?

   
  - Have you achieved the objectives?   
  - What have you learned in this summer project?
    What score do you give to yourself for this project?
  - Are there any suggestions on futher improvements
    on this probject?