Development of an evolutionary algorithm for function optimization problems with an equality constraint and its application to the inference of genetic networks

在这项研究中，我提出了一种技术，使进化算法能够解决具有多个不等式和单个不等式约束的函数优化问题。所提出的技术只是强制新生成的候选解满足等式约束。为了生成这些候选解，本研究使用了马尔可夫链蒙特卡罗方法和交叉核。该技术是为了提高遗传网络推理方法的性能而开发的。然而，我发现，即使遗传网络的推理不被定义为没有等式约束的函数优化问题，也可以推断出合理的网络。因此，本研究提出了两种不使用上述技术的遗传网络推断方法。

In this study, I proposed a technique that makes evolutionary algorithms possible to solve function optimization problems with several inequality and a single equality constraints. The proposed technique simply forces candidate solutions newly generated to satisfy the equality constraint. To generate these candidate solutions, this study uses a Markov chain Monte Carlo method and crossover kernels. This technique is developed for improving the performances of the genetic network inference methods. I found however that, even when the inference of genetic networks is not defined as a function optimization problem without having an equality constraint, reasonable networks are inferred. Thus, this study then proposed two genetic network inference methods that do not use the technique described above.