Guarded Block Transfers for Polytope Distance: Linear Convergence with Explicit Constants and Gap-Certified Screening for SVM Training
The paper introduces guarded block transfers to improve the Triangle Algorithm for solving polytope distance problems. This method aggregates multiple Mitchell–Demyanov–Malozemov transfers into one search, guarded by a comparison with the best single transfer. It achieves global linear convergence with explicit constants and includes a safe screening rule that bounds iterations based on certificate closure.