fig_ind = 1
for n in N:
x = linspace(-1,1,1001)
x_eq = array([-1.0+2.0*(i)/(n-1.0) for i in range(0,n)])
x_ch = array([cos(((2.0*i+1)/n)*0.5*pi) for i in range(0,n)])
f = zeros(n)
for i in range(0,len(f)):
f[i] = random.uniform(-1,1)
c_eq = mi.solve(n,x_eq,f); c_ch = mi.solve(n,x_ch,f)
print c_eq
p_eq = mi.horner(c_eq,x); p_ch = mi.horner(c_ch,x)
plt.figure(fig_ind)
plt.plot(x,p_eq)
plt.plot(x_eq,f,'o')
plt.title('Random equidistant points')
plt.savefig('Random_eq_N=%d.png' % n)
fig_ind+=1
plt.figure(fig_ind)
plt.plot(x,p_ch)
plt.plot(x_ch,f,'o')
x = linspace(-1,1,1001) x_eq = array([-1.0+2.0*(i)/(n-1.0) for i in range(0,n)]) x_ch = array([cos(((2.0*i+1)/n)*0.5*pi) for i in range(0,n)]) f_eq = f_book(x_eq); f_ch = f_book(x_ch) f_eq_p = zeros(len(f_eq)); f_ch_p = zeros(len(f_ch)) for i in range(0,len(f_eq)): df = rd.uniform(-0.1,0.1) f_eq_p[i] = f_eq[i] + df f_ch_p[i] = f_ch[i] + df #Find coefficients for equidistant points and Chebyshev points c_eq = mi.solve(n,x_eq,f_eq); c_ch = mi.solve(n,x_ch,f_ch) c_eq_p = mi.solve(n,x,f_eq_p) ; c_ch_p = mi.solve(n,x_ch,f_ch_p) #Evaluate p_eq and p_ch for a larger set of x-values in [-1,1] p_eq = mi.horner(c_eq,x) ; p_ch = mi.horner(c_ch,x) q_eq = mi.horner(c_eq_p,x); q_ch = mi.horner(c_ch_p,x) #Compute the errors, |p-f|, for equidistant points and Chebyshev points error_eq = abs(f_book(x)-p_eq); error_ch = abs(f_book(x)-p_ch) diff_qp_eq = abs(q_eq - p_eq); diff_qp_ch = abs(q_ch - p_ch) plt.figure(fig_ind)