Studying the influence of algorithmic parameters and instance characteristics on the performance of a multiobjective algorithm using the Promethee method.
Jochen JanssensKenneth SörensenGerrit K. JanssensPublished in: Cybern. Syst. (2019)
Keyphrases
- optimization algorithm
- cost function
- experimental evaluation
- multi objective
- improved algorithm
- dynamic programming
- significant improvement
- detection method
- computational cost
- detection algorithm
- high accuracy
- preprocessing
- objective function
- matching algorithm
- synthetic and real images
- clustering method
- parameter estimation
- levenberg marquardt
- computational complexity
- input data
- computationally efficient
- theoretical analysis
- estimation algorithm
- segmentation algorithm
- initial values
- computational efficiency
- parameters estimation
- expectation maximization
- maximum likelihood estimation
- k means
- tree structure
- parameter space
- optimization method
- convergence rate
- fine tuning
- learning algorithm
- probabilistic model
- recognition algorithm
- classification algorithm
- similarity measure
- optimal parameters
- parameter selection
- recursive least squares
- combinatorial optimization
- support vector machine svm
- energy function
- segmentation method
- maximum likelihood
- em algorithm
- search space
- neural network
- initial estimates
- uniform design