Capping methods for the automatic configuration of optimization algorithms

Automatic configuration techniques are widely and successfully used to find good parameter settings for optimization algorithms. Configuration is costly, because it necessary evaluate many configurations on different instances. For decision problems, when the objective minimize running time of algorithm, configurators implement capping methods discard poor early. Such not directly applicable optimize cost best solution found, given a predefined limit. We propose new automatic They use previous executions determine performance envelope, which cap those that do satisfy envelope conditions. integrate into irace configurator them scenarios. Our results show proposed can save from about 5% 78% effort, while finding same quality. Based computational analysis, we identify two conservative aggressive methods, an average 20% 45% respectively. also provide evidence help better available budget in scenarios with