Parallel Online Algorithms for the Bin Packing Problem

Abstract We study parallel online algorithms: For some fixed integer k , a collective of processes that perform decisions on the same sequence events forms - copy algorithm . any given time and input sequence, overall performance is determined by best individual total results. Problems this type have been considered for makespan minimization; they are also related to optimization with advice future events, i.e., number bits available in advance. Parallel algorithms interest practical scenarios which redundancy used hedging against undesired outcomes. develop Predictive Harmonic $$_3$$ 3 (PH3), relatively simple family -copy Bin Packing Problem, whose joint competitive factor converges 1.5 increasing In particular, we show $$k=6$$ k = 6 suffices guarantee 1.5714 PH3, better than 1.57829, known 1-copy Advanced while $$k=11$$ 11 achieve 1.5406, beating lower bound 1.54278 single algorithm. context advice, our approach implies 4 suffice 1.54278, considerably less previous 15 bits.