Tools › Numerical Methods › Equation Solvers › ODE and Nonlinear Equation Solver › Example
Where a circle meets an exponential curve: Newton's method for two equations with the Jacobian written out
| Tool | ODE and Nonlinear Equation Solver |
| Source | Original case, checked against scipy.optimize.fsolve and the analytic Jacobian in test_ode_nonlinear_solver.php. |
| Level | Undergraduate |
Problem
Find the point in the first quadrant where the circle x² + y² = 4 meets the curve y = eˣ − 1, by Newton's method for a system starting from (x, y) = (2, 0.5). Show the first iteration in full.
Inputs to Enter
- Problem: system of equations; x^2 + y^2 = 4 and y = exp(x) - 1
- Starting guess x = 2, y = 0.5; tolerance 1e-8
Answers
- At the guess: F₁ = 2² + 0.5² − 4 = 0.25, F₂ = 0.5 − (e² − 1) = −5.88906
- Jacobian rows (4, 1) and (−7.38906, 1); determinant 4 × 1 − 1 × (−7.38906) = 11.3891
- Step by Cramer's rule: Δx = (−5.88906 − 0.25) / 11.3891 = −0.539031, Δy = (0.25 × (−7.38906) + 5.88906 × 4) / 11.3891 = 1.90612
- After one iteration (1.4609688, 2.4061246); then (1.111307, 1.803044), (1.010054, 1.730687), (1.004186, 1.729637)
- Solution x = 1.004168738, y = 1.729637287 after 6 iterations, largest residual 8.9 × 10⁻¹⁶; measured order 2.00
Notes
Open this case in the tool: /analysis-tools/numerical-methods/equation-solvers/ode-nonlinear-solver/?try=circle-meets-curve. The two curves cross a second time in the third quadrant, near (−1.82, −0.84); a starting guess there, such as x = −2, y = −1, finds that solution instead. Newton's method returns the solution its starting guess leads to and gives no sign of the other.
More Examples for This Tool
- Lorenz equations to t = 25: adaptive Dormand-Prince RK45, the butterfly and how far chaos lets the digits be trusted
- Damped pendulum released at 172°: classical RK4 with the four slopes of the first step and a step-halving error estimate
- Colebrook friction factor at Re = 100,000 and relative roughness 0.0002 by Newton's method, iteration by iteration