A New Approximation Algorithm for Matrix Partitioning in Presence of Strongly Heterogeneous Processors.
Olivier BeaumontLionel Eyraud-DuboisThomas LambertPublished in: IPDPS (2016)
Keyphrases
- cost function
- dynamic programming
- np hard
- learning algorithm
- theoretical analysis
- experimental evaluation
- linear programming
- worst case
- similarity measure
- recognition algorithm
- convergence rate
- convex hull
- matching algorithm
- detection algorithm
- optimal solution
- significant improvement
- k means
- computational cost
- error tolerance
- closed form
- optimization algorithm
- clustering method
- evolutionary algorithm
- segmentation algorithm
- search space
- error bounds
- expectation maximization
- objective function
- polygonal approximation
- rows and columns
- simulated annealing