Genetic algorithms are surprisingly simple once you get your head around them! First I will do my best to explain them simply. Secondly, I will point you towards some references for further reading.

Remember genetic algorithms are based on natural evolution, so every part of a genetic algorithm is analogous to some aspect of natural evolution. 

First we need some way of encoding possible solutions. In natural evolution this is the DNA of an individual. In genetic algorithms however, we simply encode solutions in binary strings.

Secondly, we need a fitness function to judge how good our solutions are. In natural evolution the fitness of an individual is life it self - if you can survive long enough to have sex and have kids then you are `fit'. However, in genetic algorithms the fitness function is  defined by us. The fitness function takes an individual (a binary string) and assign a number to it corresponding to it's fitness. For a toy example, let  us assume the fitness function is how many 1's are in a binary string. For example, the individual b1= 1100 is 2, while the individual b2 = 0001 has a fitness of 1.

However the fitness function  can be anything. Perhaps, if you are creating a robot it might the distance it walks in a certain amount of time. In this case, each individual would describe the design of a robot. 

In a genetic algorithm then. First what you do is create a random population of  candidate solutions (lots of random binary strings).

Secondly for each candidate solution, you measure it's fitness using the fitness function. Then you take 2 highly fit individuals  and 'breed' them to make offspring. For example. if you breed b1 and b2 (defined above) you would chop them in half and stick the strings together to make the two kids "11001" and "0000". You keep generate kids in this way. All the kids then go into the next population. And this whole process is repeated. With our example fitness function, what you would see happen is that the new populations would tend towards the optimal solution, that is  a binary string of all ones. This is the essence of a genetic algorithm. 

See the links bellow for good books on the subject

https://wordery.com/nature-inspired-metaheuristic-algorithms-xin-she-yang-9781905986286?currency=GBP>rck=bHFnRHdUeGlGUHJEUGg0RHVuRnRuLy84U01CeVJHSFFPdkJUdnJYNE85VGpQVDMzK0VrRVA2UkR1enREak80cGJnNXdrdHlHTERBV3JVZERpMWE2aHc9PQ&gclid=CjwKEAiA04S3BRCYteOr6b-roSUSJABE1-6BP7GSD1UNHKmctqGg2EE0DlejJFb9pUdx-unxE1IEnhoCO-rw_wcB

https://wordery.com/introduction-to-genetic-algorithms-sn-sivanandam-9783540731894?currency=GBP>rck=dnVxdzhoUUc1T3NlQmhiTjArZlpwbkpFa0RXU1J0Znp2WU5RYW9TbWIyZC9RM1JDYmFRYlJXajVPS0Nsa2M5SzhMZEd1a2pWSzNVRDhQRkpXVnBST0E9PQ&gclid=CjwKEAiA04S3BRCYteOr6b-roSUSJABE1-6B3F1BxeCWMO9v-ZUI3dQieuEQ-bE59L4FsxEMXtF14xoCIG7w_wcB

In addition to Jack's comprehensive explanation, concisely I can say:

GA is a simple, but powerful technique to obtain a set of local or global solutions in different problems. Beside learning this technique which is easy, you need to be expert in your problem area. If you see the vast applications of EAs (Evolutionary Algorithms), you understand that the most important part is to know how to model your problem in such a way to fit on EA. 

However, my suggestion is to read related papers to your problem area which use GA or any type of EA, to get an idea about modeling your problem. Also you need to check examples of GA on different applications. 

Regards, 

I also agree with Mr. Jack's and Reza's explanation. I also suggest you first focus on your problem and read related papers and findout examples of GA. there are lot of GA examples are available on net.

You can also try to read my paper for getting some basic knowledge:

https://www.researchgate.net/publication/257018767_Dynamic_optimization_of_inside_temperature_of_Zero_Energy_Cool_Chamber_for_storing_fruits_and_vegetables_using_neural_networks_and_genetic_algorithms

Article Dynamic optimization of inside temperature of Zero Energy Co...

Thank you so much, Jack Mckay Fletcher, Reza Javanmard Alitappch, Md.Parvez Islam...for sharing those valuable information. I can now understand the fundamentals/basics and purpose of  GA's now but still, i am stuck with the fitness function!

Very less papers exists related to my area of interest and those papers haven't justified the fitness function formulae. Its, as if, it just exist as a function; in those papers. So, it would be helpful if a specific problem with GA as a probable solution is explained, with an extrapolated explanation for the existence of fitness function for that problem.

Meantime i will go through those books and a few other papers for sure.

Regards,

i am getting the concepts you (Jack Mckay Fletcher, Reza Javanmard Alitappch, Md.Parvez Islam) all were trying to convey and it did help me in looking for answers in right places. Right now i am stuck with framing a polynomial equation for initial population part. It would be helpful if anybody can give me advice regarding the same.

Regards,