first you need to define your data as cromosom in GA, as in my research on Vehicle routing problem define base route and its distance,
second create your population example in mine create initial distribution route and code it,
third create fitness and its probability value, fitness will be total of distance divide by value of each vehicle route, accumulate all as total fit, as probability fitnes each route divide by total fitness
fourth, selection code using its route and probability
fifth, crossover process.
sixth, will be its mutation
last will be result from mutation as decision result.