Printable View
My question is basically the primary stage prior to coding. I am trying to do a problem of solid waste collection vehicle routing using Genetic algorithm in C++. So here raises the question why should I use GA. How do I defense this question? My main logic is solid waste collection problem is a non linear problem and for solving it I chose GA. So can it be like I can form a model to prove its non linearity? If its modeling is possible what things do I need to focus?
Note to mods: this is not a C++ problem, and this thread probably belongs in the "General Discussions" or "Tech Board" forum.
To answer the question, given what little there is of your problem description, you are seeking to find a path through some set of points that is optimal in some measure. Such problems are typically NP-complete, which means there is no a specific algorithm for solving such problems analytically. Evolutionary search techniques are typically used to find possible solutions to problems that are not solvable analytically. The genetic algorithm is just one such technique that is often used in practice.
You would need to do more analysis of your problem space before you can actually defend using a GA to find a solution. GA is better suited to problems in which an existing feasible (or starting) solution can be improved with small parameters changes, and where the cost function for new solutions is not particularly demanding computationally. If your problem is not like that, then other approaches may be more suitable. Then it comes down to a trade-off between your effort (to analyse the problem space) versus being able to run with a particular technique. In practice, the GA is often used as a convenient fallback when people can't be bothered analysing their problem space, so is often applied to problems for which it may not be the best approach.
Moved, and given a decent title.