Difference of Genetic Algorithm and Constraint Programming?

Question

I hope someone would shed some light on me about this topic. If by any chance this is considered a stupid question to ask, I'd gladly remove this question right away.

I am designing a course timetabling system and by researching, I stumbled upon GA and Constraint Programming as approaches in solving my problem. However, I didn't quite understand the differences between the two and what are the advantages of one over the other. I hope someone would explain this to me in layman's term or direct me to a site with this topic.

Thanks in advance!

Best regards.

Answer 1

Here's how I see the family of optimization algorithms:

• Exact methods: brute force, branch and bound
• Constraint Programming (terrible name): tries reducing the domain set
• Linear Programming et al: simplex, ...
• Metaheuristics:
  • Local Search: Tabu Search, Simulated Annealing, Late Acceptance, ...
  • Population based algorithms: genetic algorithms, swarm optimization, ...

For the use case course timetabling specifically, the ITC2007 research competition clearly showed that Local Search is king. Genetic algorithms were consistently slightly inferior and Constraint Programming was useless due to scalability issues. Your mileage may very as 2007 is some time ago.