Genetic Algorithm with spatial data

Hi everyone, I am wanting to use genetic algorithm for road alignment optimisation. I was wondering if any one would have any tips for how to structure the input data. I want to use two variables as cost constraints but have them as multiple objectives to see the relationship between the two variables. My main question is what should the input data look like. Should I simply have my network of nodes and arcs and then a cost matrix. Or can I read the costs off an attribute in my network?

Any help with integrating spatial data into Genetic algorithms would be appreciated.