在計算機科學中最常見到的資料結構便是圖形，圖形基本上是由節點和邊所組合而成。圖型結構往往隨著龐大的資料量而變的相當複雜，要如何透過電腦的輔助，才可以將複雜的圖形結構繪製成概念清楚的圖形。 在[Hsieh05]中，對於二維空間上直角多邊形內繪製樹狀圖的問題，提出一個以基因演算法為基礎的樹狀圖繪製演算法，透過階層式的繪製方式，將樹狀圖收縮後再逐層展開來做繪製。 然而[Hsieh05]所設計的繪圖演算法存在不少的缺點，當遇到結構龐大的樹狀圖執行時間過長，突變和交配運算的設計不夠恰當，且美學規則設計並不能明確地挑出較佳的樹狀圖。此篇論文，主要是針對[Hsieh05]所提出的繪圖演算法作改良，我們藉由物種初始狀態的改良來幫助改善過長的執行時間，且提出新的交配和突變方式，並在美學規則加入新的概念。 The data structure seen most frequently in the computer science is graph, and the graph is composed of nodes and edges. The graph often becomes complicated with a large number of data, so we must change the complicated graph into the graph that is clear and easy to understand by using computer. In the [Hsieh05], it provides a drawing algorithm which is based on the genetic algorithm to draw a tree in the rectilinear polygon on the two-dimentional space. By using Hierarchical layout, the drawing algorithm contracts a tree and expands it hierarchically. However, the drawing algorithm has many disadvantages. It spends a lot of time drawing a complicated tree; designs for crossover and mutation are not good enough, and aesthetic criteria can’t pick out a proper tree. This paper focuses on improving the drawing algorithm. We improve the initial state of species to prevent lasting execution time, provide new designs for crossover and mutation, and join a new concept in aesthetic criteria.