在計算神經科學中，神經迴路模型的發展與調整通常依靠經驗以及簡單的最佳化方法。因此，為了能夠發展出應用於神經迴路且有效率的最佳化演算法，我們設計了一套演化的機制，讓系統自行尋找及演化為最佳的參數，甚至發展出全新的神經迴路。為完成這個計畫，我們利用基因演算法（Genetic Algorithm）建構此演化機制，並應用在神經迴路與環境的模擬系統（Hanitu）。我們證實基因演算法能夠有效的作用在神經迴路的演化機制中，除了能改變神經迴路的模型參數，還能演化神經迴路的結構。過程中，特別的是我們設計了一套創新的基因編碼方式，利用這套基因編碼方式能夠紀錄神經迴路的結構與所有參數，演化出有不同神經元與突觸的神經迴路，並且在兩百世代之後能夠使評估值上升為原評估值的六倍。我們的研究結果證實此演化機制能夠使神經迴路的模型參數最佳化，同時演化出有不同結構、全新的神經迴路。

In computational neuroscience, we often used our experience or conventional optimization methods to optimize the neural circuit models. To develop a more efficient optimization algorithm, it is necessary to design a mechanism in which the neural circuit could evolve and optimize itself. To this end, we developed a method based on Genetic Algorithm and applied our method on the Hanitu System, which is a neural network simulation platform. We demonstrated that Genetic Algorithm worked well on the evolution of neural circuit models. In addition to the simple optimization method which only optimized the model parameters of the neural circuit models, we further developed an advanced evolution mechanism with which the nervous systems were able to develop new neurons and new links. Specifically, we proposed a novel gene coding scheme which encoded all parameters and the circuit structure in a hierarchical fashion. Using this novel mechanism of evolution, we could obtain the fitness level that was about six times higher than the original one in less than 200 generations. The system also evolved new circuits with different connections and parameters. In conclusion, the result demonstrated that our Genetic Algorithm-based evolution mechanism could not only optimize the parameters of the existing circuits but also create new circuit architectures, or new species.

Lin, F.-T. (2005). Evolutionary computation part 1: genetic algorithms and their three
applications. Journal of Taiwan Intelligent Technologies and Applied Statistics, 3(1), 1-28.
Lin, F.-T. (2005). Evolutionary computation part 2: genetic algorithms and their three
applications. Journal of Taiwan Intelligent Technologies and Applied Statistics, 3(1), 29-56.
Weise, T. (2009). Global optimization algorithms-theory and application. Self-
published, 2.
Tsai, S.-Y. (2012). How worms find their foods? Neural hysteresis, working memory
and attractors in a simple neural circuit. (Master’s thesis, National Tsing Hua
University, Department of Physics).
Hsieh, F.-K. (2016). Computational neuroethology: Developing a virtual worm
system and its energy homeostasis mechanism. (Master’s thesis, National Tsing
Hua University, Institute of Systems Neuroscience).
Jaddi, N. S., Abdullah, S., & Hamdan, A. R. (2016). A solution representation of
genetic algorithm for neural network weights and structure. Information Processing Letters, 116(1), 22-25.
Haferlach, T., Wessnitzer, J., Mangan, M., & Webb, B. (2007). Evolving a neural
model of insect path integration. Adaptive Behavior, 15(3), 273-287.
Homem, C. C., & Knoblich, J. A. (2012). Drosophila neuroblasts: a model for stem
cell biology. Development, 139(23), 4297-4310.
Hopfield, J. J., & Tank, D. W. (1985). “Neural” computation of decisions in
optimization problems. Biological cybernetics, 52(3), 141-152.
Zaabab, A. H., Zhang, Q. J., & Nakhla, M. (1995). A neural network modeling
approach to circuit optimization and statistical design. IEEE Transactions on
Microwave Theory and Techniques, 43(6), 1349-1358.
Krauth, W., & Mézard, M. (1987). Learning algorithms with optimal stability in
neural networks. Journal of Physics A: Mathematical and General, 20(11), L745.
Srinivas, M., & Patnaik, L. M. (1994). Adaptive probabilities of crossover and
mutation in genetic algorithms. IEEE Transactions on Systems, Man, and
Cybernetics, 24(4), 656-667.
Smith, K. A. (1999). Neural networks for combinatorial optimization: a review of
more than a decade of research. INFORMS Journal on Computing, 11(1), 15-34.
Xia, Y., Leung, H., & Wang, J. (2002). A projection neural network and its
application to constrained optimization problems. IEEE Transactions on Circuits
and Systems I: Fundamental Theory and Applications, 49(4), 447-458.
Yao, X. (1999). Evolving artificial neural networks. Proceedings of the IEEE, 87(9),
1423-1447.
Adeli, H., & Park, H. S. (1995). A neural dynamics model for structural
optimization—theory. Computers & structures, 57(3), 383-390.
Tagliarini, G. A., Christ, J. F., & Page, E. W. (1991). Optimization using neural
networks. IEEE transactions on computers, 40(12), 1347-1358.
Sims, K. (1994, July). Evolving virtual creatures. In Proceedings of the 21st annual
conference on Computer graphics and interactive techniques (pp. 15-22). ACM.
Alpaydin, G., Balkir, S., & Dundar, G. (2003). An evolutionary approach to automatic
synthesis of high-performance analog integrated circuits. IEEE Transactions on
Evolutionary Computation, 7(3), 240-252.
Kermani, B. G., Schiffman, S. S., & Nagle, H. T. (1999). Using neural networks and
genetic algorithms to enhance performance in an electronic nose. IEEE
Transactions on Biomedical Engineering, 46(4), 429-439.
Ludermir, T. B., Yamazaki, A., & Zanchettin, C. (2006). An optimization
methodology for neural network weights and architectures. IEEE Transactions on Neural Networks, 17(6), 1452-1459.
Jones, A. J. (1993). Genetic algorithms and their applications to the design of neural
networks. Neural Computing & Applications, 1(1), 32-45.
Chen, H. (1995). Machine learning for information retrieval: neural networks,
symbolic learning, and genetic algorithms. Journal of the Association for
Information Science and Technology, 46(3), 194-216.