Hill-climbing and branch-and-bound algorithms for exact and approximate inference in credal networks.
Andrés CanoManuel Gómez-OlmedoSerafín MoralJoaquín AbellánPublished in: Int. J. Approx. Reason. (2007)
Keyphrases
- hill climbing
- exact and approximate
- branch and bound algorithm
- branch and bound
- credal networks
- search space
- lower bound
- search algorithm
- bayesian networks
- maximum entropy
- set valued
- upper bound
- simulated annealing
- genetic algorithm ga
- search strategy
- optimal solution
- search tree
- combinatorial optimization
- search procedure
- np hard
- lower and upper bounds
- exhaustive search
- objective function
- metaheuristic
- heuristic search
- query evaluation
- probabilistic inference
- fitness function
- string matching
- constraint satisfaction problems
- dynamic programming
- search methods
- directed acyclic graph
- conditional random fields
- variational inequalities
- cost function