| Popis: |
Many long term robot exploration domains have sparse fitness functions that make it hard for agents to learn and adapt. This work introduces Adaptive Multi-Fitness Learning (A-MFL), which augments the structure of Multi-Fitness Learning (MFL) [7] by injecting new behaviors into the agents as the environment changes. A-MFL not only improves system performance in dynamic environments, but also avoids undesirable, unforeseen side-effects of new behaviors. On a multi-robot coordination problem, A-MFL provides up to 90% improvement over MFL and 100% over a one-step evolutionary approach. |
