Algorithm Research And Software Implementation For Virtual Plant Growth Based On L-Systems

Virtual plant growth studies how to simulate the process of plant development in computers. With the rapid development of computer technology, it has become a hot research area in computer science, and been applied into many practical fields, such as agriculture, education, gardening, business and entertainment at present.L-System is a commonly used mathematical methodology to generate fractals and realistic model plants. In this dissertation we study some basic algorithms of L-System, and develop an extendible, context-sensitive, bracketed, and stochastic L-System, called ParaLSys. Besides, in order to demonstrate the availability of ParaLSys, the data exchange scheme between a classic open souce L-System implementation, LParser, and ParaLSys is studied.In Chapter 1, the research background is introduced, and relative research work is reviewed.According to the modeling ways, L-System is classified into several categories, such as DOL-System (deterministic and context-free L-System), stochastic L-System, context-sensitive L-System and parametric L-System. In Chapter 2, these L-Systems will be discussed respectively in detail.In Chapter 3, the design and implementation issues of ParaLSys are illustrated. It is built on the OpenGL graphical libarary and the QT UI library, and integrated with a self-defined descriptive language. Five modules are contained in ParaLSys, their functionalities are to edit the description texts, to setup mesh, to setup alphabet, to evolve the system, and to view the results, respectively. In order to improve the simulation efficiency, the algorithms are built on binary trees rather than traditional sequent arrays. Experiment shows that better timing and spacing performance is achieved with the new data structure.In Chapter 4, performance of LParser and ParaLSys is compared. Implementation of data exchange scheme between both systems is discussed, where the focus is how to translate commands in LParser, such as the turtle orientation, turtle moving, inc/dec, into the descriptive language in ParaLSys. Two examples are presented to prove that major functionalities of LParser could be imitated by ParaLSys. Since LParser has been successfully applied into many important areas, it is expected that ParaLSys could also be used to resolve the same problems.Finally, in Chapter 5, the dissertation is summarized and some proposals for future work are given.