Ranking and selection for pairwise comparison

In many real-world applications, designs can only be evaluated pairwise, relative to each other. Nevertheless, in the simulation literature, almost all ranking and selection procedures are developed based on individual performances of design. This research considers statistical problem when design performance simulated pairwise. We formulate this new using optimal computing budget allocation approach derive asymptotic optimality condition some approximations. The numerical study indicates that our reduce number simulations required confidently identify best