dannyi96 / pygenetic Goto Github PK

Currenty(with distinct crossover and swap mutation), n Queens problem doesn't get solved for even small numbers like n=16
We need to include other methods or figure other better GA parameter configuration to solve problem for larger values of n

Need to implement ga which is currently a global variable as a cookie to support multiple users accessing it at the same time

Download equivalent code as a python file on pressing the form submit button

Need to Redesign selection handlers to work based on fitness_type max, min and equal

Indicated by Green/Red Colour

Need to have a final review of the entire code and work on publishing it on pypi

Handle different fitness types - MAXIMUM, MINIMUM ( with and without threshold), EQUAL

Implementation of GA Optimisations like adaptive mutation

Need to create a good README section

One of the outputs of the crossovers for inputs ([1, 0, 7, 2, 6, 4, 3, 5], [4, 5, 3, 6, 0, 1, 2, 7]) remained ([1, 0, 7, 2, 6, 4, 3, 5], [4, 5, 3, 6, 0, 1, 2, 7]).

Need to improve the appearance of the textarea boxes

Create some ipython notebook files of Step by Step Code Examples with explanation ( nQueens and TSP )

Need to add scrolling to the bottom of page effect as and when each generation data is received.

Convert user inputs from UI to equivalent python code

Hello, having a little trouble using your library so I will probably fork it.

Problem

Pygenetic is not using a version of matplotlib which can be installedin python 3.8
Version 2.2.2 of matplotlib is not available

To reproduce:

Make or access a python 3.8 install.
Attempt to install pygenetic, or even matplotlib 2.2.2
Get told it is not possible, as matplotlib 2.2.2 cannot be found.

I will open a pull request once I know updating to a newer version will not break anything.

Need to have a population control parameter in GAEngine
If set to true - always ensure that population in each generation is same ( in case it isn't , it selects only needed elements from top of the list )
If set to false - Population can take any number of members. (i.e. - can always be constantly increasing ). User can optionally set a UPPER_LIMIT here. Useful for pySpark Evolutions

Handling of multiple crossovers and mutation handlers with different probabilities

Need to include various statistics of GA to visualize best/avg/max/min fitness over different generations, etc

2D Chromosome Factory which are popularly used in many problems like sudoku solving

Explore and implement some ML technique using GA

Implementing polling to get each generation data on the UI immediately as soon as it is computed.

We may need to go for running it in a cluster. Also try to optimise the various RDD actions to a minimum

https://stackoverflow.com/questions/40502185/spark-program-takes-a-really-long-time-to-complete-execution

Roulette Wheel Selection
Stochastic Universal Sampling (SUS)
Tournament Selection
Rank Selection
Random Selection

Store best fitness throughout all generations ( applicable to all selection types )

May fail due to the use of the list p which contains relative fitnesses

Implement Jquery functionality to minus(-) button

Need to integrate TravisCI in our project

We have to modify the formula a bit or use a different adaptive method to see a noticeable difference

mutation_rates [0.5859037956712174, 0.5890376972849537, 0.5903909891318503, 0.5916622332897328, 0.5910757976952127, 0.5911468264032488, 0.5913217861004001, 0.5912461103370071, 0.5916085995081001, 0.5918987263324281, 0.5922810038663388, 0.5924175437299303, 0.5919739575621374, 0.5920402257273272, 0.5917604537840548, 0.5924933255524385, 0.5923069473804097, 0.5923912569728325, 0.5926584181942748, 0.5922699332103435]

So users can set different optimisations

• adaptive_mutation (true/false)
• injection of half of fame (true/false)
• stop iterations (true/false)

Aim: Incorporate different map reduce architectures for GA Evolution

Implementation of a Simple Basic High Level GA Api

Need to explore more application/examples of GA and implement it using our API

Need to allow option for custom crossover, mutation, selection and evolution from UI

Need to change evolution using pyspark class since it supports only limited selection - best, worst and equal to selection criteria. Need to ignore other given selection criteria.
Need to change the interfaces as well

Termination of iterations when fitness type is EQUAL for flask app

Compare different GAs executed based on

• Usage of pyspark vs normal execution
• pySpark Execution: comparision of different configs( number of masters/workers)
• Usage of different selection, crossover, mutation methods
• Usage of different optimisation techniques vs without using them

And completion of final report

Chromosome Factory testing

• Testing Abstract class nature of ChromosomeFactory
• Testing ChromosomeRangeFactory: Correct chromosomes given valid (data_type, noOfGenes, minValue, maxValue, duplicates)
• Testing ChromosomeRangeFactory: Throws exception for invalid (data_type, noOfGenes, minValue, maxValue, duplicates). Invalid includes cases like max being lesser than min, wrong datatypes for different inputs, inability to convert to int, etc
• Testing ChromosomeRegexFactory: Correct chromosomes given valid (data_type,noOfGenes,pattern).
• Testing ChromosomeRegexFactory: Throws exception for invalid (data_type,noOfGenes,pattern). Invalid includes cases like invalid regex, wrong datatypes for different inputs, inability to convert to int, etc

Utils testing

• Testing Fitness Utils which works for correct inputs and throws Exception for invalid exceptions
• Testing CrossoverHandlers Utils which works for correct inputs and throws Exception for invalid exceptions. Need artificial data (since randomness is involved may need to include all possible outputs )
• Testing MutationHandlers Utils which works for correct inputs and throws Exception for invalid exceptions. Need artificial data (since randomness is involved may need to include all possible outputs )
• Testing SelectionHandlers Utils which works for correct inputs and throws Exception for invalid exceptions

Population Testing

• Test if initialisation is done correctly and proper members are created based on the ChromosomeFactory given as input
• Test population members before and after selection, crossover, mutation

Statistics Testing

• Test if valid best, worst, avg, diversity, adaptive mutation are stored after each iteration
• Test all the methods of Statistics class
• Test if exception is thrown on invalid inputs
• Test if correct graph is created

Evolution Testing

• Test if evolution works correctly during different iterations ( ie - correct selection, crossover, mutation takes place. Need artificial data )
• Test if exceptions are thrown for incorrect

GAEngine Testing

• Test all the basic methods (setters,etc) of GAEngine class (valid , invalid inputs)
• Test best_fitness during iterations

Making the user interface of main page of our web application (to get details of the GA to be executed)