My question is about how heuristic/artificial intelligence, such as swarm intelligence, game theory, can be in the approach of multi objective decision making? Can any body guide me?
Fábio Lobato
Hi Lezli,
First thing, you just need multi objective approaches if your evaluation measures are conflicting. We have three approaches to handle MODM: transform it in a single objective approach using a weighted equation, lexicographic prioritization and Pareto exploration methods. The choice depends on many factors.
I hope that answer can help you a little bit.
Regards,
Fábio
1 votes
0 thanks
Paweł Kubik
Get rid of the multiple criteria by applying some function (e.g the linear form). It would be best to use a domain knowledge of given problem. Many multiple criteria problems are actually single criteria problems. Replace some of the objectives with constraints. Consider the knapsack problem. You can define it as multi criteria problem of minimizing the weight and maximizing the cost., however you can usually replace the weight criteria with some fixed constraint like the capacity of your actual knapsack. You can try to use a swarm algorithm (e.g. evolutionary algorithm) in order to find the Pareto frontier, which is a set of optimal solution. This is the most difficult solution and requires to put a lot of effort into finding the proper algorithm parameters.
There are actually 3 ways to approach this problem:
https://en.wikipedia.org/wiki/Linear_form
https://en.wikipedia.org/wiki/Knapsack_problem
https://en.wikipedia.org/wiki/Pareto_efficiency
1 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence
More Lezli Brown's questions See All
Similar questions and discussions