進化式演算法早已經受到廣泛的應用，這類方法擁有的特性非常適合用來處理一些問題，例如黑箱與NP-hard (Non-deterministic Polynomial-time hard)的問題. 然而可惜的是，這類方法已經被證實屬於耗時(time consuming)的方法，這與實務上的需求是有差異的，例如: 工廠不可能花費一整天的時間等待一個最佳的排程、企業不可能耗費1個月進行資料的精簡。因此，縮短這類演算法的運算時間是極為重要的。 本研究提出了一個新穎的最佳化演算法名為快速基因演算法(Efficient-Genetic Algorithm, EGA), 透過增加更多的生物學的原理來改善演算法的演化效率。 換句話說，在有限的資源下，生物經過了長時間的演化，找出了最有效率的演化以及存活方法，這是值得參考的。 本文透過兩個主要的實驗來驗證EGA的有效性，在第一個實驗中，針對EGA測試了兩個常見的最佳化問題。此外，本研究也納入了兩個知名的演算法分別是基因演算法(Genetic algorithm, GA)以及其改良的變種版本免疫演算法(Immunity algorithm, IA)。然而在第二個實驗中，為了貼近實務的需求，本研究採用了四個高維度的資料集進行資料縮減的實驗，並與四個知名的方法進行比較，四種方法分別為IB3、DROP3、ICF 還有GA。 ;Evolutionary computations have been widely used in many real word problems. In particular, evolutionary algorithms can be considered as the global optimization methods with a met heuristic or stochastic optimization character and they are widely applied for black box problems (no derivatives known) and non-deterministic polynomial-time hard problems (NP-hard), often in the context of expensive optimization. However, their computational complexities are very high leading to the major limitation in practice. In this dissertation, we introduce a novel Efficient-Genetic Algorithm (EGA), which fits “biological evolution” into the evolutionary process. In other words, after long-term evolution, individuals find the most efficient way to allocate resources and evolve. There are two experiments to validate the EGA. The first experimental study is based on a scheduling problem, and two state-of-the-art algorithms including Genetic algorithm (GA) and Immunity algorithm (IA) are compared with EGA. The second one focuses on the data reduction problem where four very high dimensional datasets are used. In addition, four state-of-the-art algorithms including IB3, DROP3, ICF, and GA are compared with EGA.
