课程设计报告_四阶Runge_Kutta方法.doc

PAGE  PAGE 12 《计算机数值方法》 课程设计报告 题目四阶Runge-Kutta方法学生姓名班级计科12学号 成绩指导教师 延安大学计算机学院 2014年9月1日 目 录 摘要·······································································5 问题重述··································································5 方法原理及实现··························································5 计算公式或算法··························································5 Matlab程序·······························································6 测试数据及结果··························································6 结果分析·································································10 八、方法改进··································································10 九、心得体会··································································10 十、参考文献··································································10 摘要 本课程设计主要内容是用四阶Runge-Kutta方法解决常微分方程组初值问题的数值解法，首先分析题目内容和要求，然后使用Matlab编写程序计算结果并绘图，最后对计算结果进行分析并得出结论。 问题描述 在计算机上实现用四阶Runge-Kutta求一阶常微分方程初值问题 的数值解，并利用最后绘制的图形直观分析近似解与准确解之间的比较。 三、方法原理及实现 龙格-库塔(Runge-Kutta)方法是一种在工程上应用广泛的高精度单步算法。由于此算法精度高，采取措施对误差进行抑制，所以其实现原理也较复杂。该算法是构建在数学支持的基础之上的。龙格库塔方法的理论基础来源于泰勒公式和使用斜率近似表达微分，它在积分区间多预计算出几个点的斜率，然后进行加权平均，用做下一点的依据，从而构造出了精度更高的数值积分计算方法。如果预先求两个点的斜率就是二阶龙格库塔法，如果预先取四个点就是四阶龙格库塔法。 经典的方法是一个四阶的方法，它的计算公式是： 四、计算公式或算法 输入（编写或调用计算的函数文件）， 3．For End 4．输出 五、Matlab 程序 x=[a:h:b]; y(1)=y1; n=(b-a)/h+1; for i=2:n fk1=f(x(i-1),y(i-1)); fk2=f(x(i-1)+h/2,y(i-1)+fk1*h/2); fk3=f(x(i-1)+h/2,y(i-1)+fk2*h/2); fk4=f(x(i-1)+h,y(i-1)+fk3*h); y(i)=y(i-1)+h*(fk1+2*fk2+2*fk3+fk4)/6; end y 六、测试数据及结果 用调试好的程序解决如下问题： 应用经典的四阶Runge-Kutta方法解初值问题 取 步骤一：编写函数具体程序. 1.求解解析解程序： dsolve(Dy=(y^2+y)/t,y(1)=-2,t) 结果： 2.综合编写程序如下： a=1; b=3; h=0.5; y(1)=-2; x(1)=a; n=(b-a)/h+1; yy(1)=-2; for i=2:n k1=(y(i-1)^2+y(i-1))/x(i-1); k2=((y(i-1)+h*k1/2)^2+(y(i-1)+h*k1/2))/(x(i-1)+h/2); k3=((y(i-1)+h*k2/2)^2+(y(i-1)+h*k2/2))/(x(i-1)+h/2); k4=((y(i-1)+h*k3)^2+(y(i-1)+h*k3))/(x(i-1)+h); y