Adaptive Structural Hyper-Parameter Configuration by Q-Learning
This addresses hyper-parameter tuning for evolutionary algorithms, which is an incremental improvement as it applies a known method (Q-learning) to a specific bottleneck in a domain-specific context.
The paper tackles the problem of tuning structural hyper-parameters in evolutionary algorithms by modeling it as a reinforcement learning problem, specifically using Q-learning to control computational resource allocation in a CEC 2018 winner algorithm, and shows favorable results against that algorithm on CEC 2018 test functions.
Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.