ABSTRACT
In this study, a neuroevolution strategy using Multi-Agent Incorporated Hierarchical Ensemble Model (MAIHEM) inspired by human incorporated company structure is proposed. It utilizes the hierarchical structure to ensemble modules of entities into firms to preserve complex structure at lower level, and at higher level incorporates firms into departments to facilitate multiple objectives. The corporate level structure from the top guides and reviews their overall performance. The ensemble structure not only compete within their own ranks, but also cooperate and swap/merger underlying units for fast adaptation without compromising their existing structures. The preliminary result with multi-constrained music melody generation shows this strategy can not only solve complex multi-objective tasks steadily but also preserve diversity in the population.
- F. P. Such, V. Madhavan, E. Conti, J. Lehman, K. O. Stanley, J. Clune. 2017. Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning. arXiv preprint arXiv:1712.06567.Google Scholar
- M. Lin, Q. Chen, S. Yan. 2013. Network in network. arXiv preprint arXiv:1312.4400.Google Scholar
- K.-W. Su. 2012. Study on Retrieving Subjective Knowledge from Music Using Adaptive Cloud-Based Structure. Master's thesis. National Taiwan University of Science and Technology (NTUST), Taipei, Taiwan.Google Scholar
- J. Lehman, K. O. Stanley. 2011. Abandoning objectives: Evolution through the search for novelty alone. Evolutionary computation, 19(2), 189--223. Google ScholarDigital Library
- MAIHEM melody generation samples. 2018. https://goo.gl/GmGVLq Accessed: 2018- 02- 06.Google Scholar
Index Terms
- A neuroevolution strategy using multi-agent incorporated hierarchical ensemble model
Recommendations
Cultivating Diversity: A Comparison of Diversity Objectives in Neuroevolution
Applications of Evolutionary ComputationAbstractInspired by biological evolution’s ability to produce complex and intelligent beings, neuroevolution utilizes evolutionary algorithms for optimizing the connection weights and structure of artificial neural networks. With evolutionary algorithms ...
Improving Distributed Neuroevolution Using Island Extinction and Repopulation
Applications of Evolutionary ComputationAbstractNeuroevolution commonly uses speciation strategies to better explore the search space of neural network architectures. One such speciation strategy is the use of islands, which are also popular in improving the performance of distributed ...
Application of the Hierarchic Memetic Strategy HMS in Neuroevolution
Computational Science – ICCS 2022AbstractQuite recently some noteworthy papers appeared showing classes of deep neural network (DNN) training tasks where rather simple one-population evolutionary algorithms (EA) found better solutions than gradient-based optimization methods. However, it ...
Comments