Multi-Step Methods

A multi-step method computes wi+1w_{i+1} from several previous approximations, not just wiw_i.

An mm-step method for the IVP y′(t)=f(t,y)y'(t) = f(t, y) has a difference equation

wi+1=am−1wi+am−2wi−1+⋯+a0wi+1−m+h[bmf(ti+1,wi+1)+bm−1f(ti,wi)+⋯+b0f(ti+1−m,wi+1−m)]\begin{aligned} w_{i+1} = {}& a_{m-1} w_i + a_{m-2} w_{i-1} + \cdots + a_0 w_{i+1-m} \\ & + h\big[ b_m f(t_{i+1}, w_{i+1}) + b_{m-1} f(t_i, w_i) + \cdots + b_0 f(t_{i+1-m}, w_{i+1-m}) \big] \end{aligned}

for i=m−1,m,…,N−1i = m-1, m, \ldots, N-1, with constants a0,…,am−1a_0, \ldots, a_{m-1} and b0,…,bmb_0, \ldots, b_m, and h=b−aNh = \dfrac{b-a}{N}.

Starting values w0=αw_0 = \alpha and w1,…,wm−1w_1, \ldots, w_{m-1} come from a one-step method such as Runge-Kutta.

Types

Explicit Method

bm=0b_m = 0. wi+1w_{i+1} is given directly by previously determined values.

Implicit Method

bm≠0b_m \neq 0. wi+1w_{i+1} appears on both sides and must be solved for.

Adams Methods

Integrating y′=f(t,y)y' = f(t, y) over [ti,ti+1][t_i, t_{i+1}]

y(ti+1)=y(ti)+∫titi+1f(t,y(t)) dty(t_{i+1}) = y(t_i) + \int_{t_i}^{t_{i+1}} f(t, y(t))\, \text{d}t

f(t,y(t))f(t, y(t)) is replaced by a polynomial through the mm most recent points, then integrated.

Below, fk=f(tk,wk)f_k = f(t_k, w_k) and μi\mu_i lies in the interval spanned by the points used. τi+1(h)\tau_{i+1}(h) is the local truncation error per unit step. Substitute the exact values y(tk)y(t_k) into the method and divide the remaining residual by hh. It measures the error introduced at step i+1i+1 when all preceding values are exact.

Adams-Bashforth Explicit Methods

2-step, with τi+1(h)=512y′′′(μi)h2\tau_{i+1}(h) = \dfrac{5}{12} y'''(\mu_i) h^2

wi+1=wi+h2[3fi−fi−1]w_{i+1} = w_i + \frac{h}{2}\left[ 3 f_i - f_{i-1} \right]

3-step, with τi+1(h)=38y(4)(μi)h3\tau_{i+1}(h) = \dfrac{3}{8} y^{(4)}(\mu_i) h^3

wi+1=wi+h12[23fi−16fi−1+5fi−2]w_{i+1} = w_i + \frac{h}{12}\left[ 23 f_i - 16 f_{i-1} + 5 f_{i-2} \right]

4-step, with τi+1(h)=251720y(5)(μi)h4\tau_{i+1}(h) = \dfrac{251}{720} y^{(5)}(\mu_i) h^4

wi+1=wi+h24[55fi−59fi−1+37fi−2−9fi−3]w_{i+1} = w_i + \frac{h}{24}\left[ 55 f_i - 59 f_{i-1} + 37 f_{i-2} - 9 f_{i-3} \right]

5-step, with τi+1(h)=95288y(6)(μi)h5\tau_{i+1}(h) = \dfrac{95}{288} y^{(6)}(\mu_i) h^5

wi+1=wi+h720[1901fi−2774fi−1+2616fi−2−1274fi−3+251fi−4]w_{i+1} = w_i + \frac{h}{720}\left[ 1901 f_i - 2774 f_{i-1} + 2616 f_{i-2} - 1274 f_{i-3} + 251 f_{i-4} \right]

Adams-Moulton Implicit Methods

2-step, with τi+1(h)=−124y(4)(μi)h3\tau_{i+1}(h) = -\dfrac{1}{24} y^{(4)}(\mu_i) h^3

wi+1=wi+h12[5fi+1+8fi−fi−1]w_{i+1} = w_i + \frac{h}{12}\left[ 5 f_{i+1} + 8 f_i - f_{i-1} \right]

3-step, with τi+1(h)=−19720y(5)(μi)h4\tau_{i+1}(h) = -\dfrac{19}{720} y^{(5)}(\mu_i) h^4

wi+1=wi+h24[9fi+1+19fi−5fi−1+fi−2]w_{i+1} = w_i + \frac{h}{24}\left[ 9 f_{i+1} + 19 f_i - 5 f_{i-1} + f_{i-2} \right]

4-step, with τi+1(h)=−3160y(6)(μi)h5\tau_{i+1}(h) = -\dfrac{3}{160} y^{(6)}(\mu_i) h^5

wi+1=wi+h720[251fi+1+646fi−264fi−1+106fi−2−19fi−3]w_{i+1} = w_i + \frac{h}{720}\left[ 251 f_{i+1} + 646 f_i - 264 f_{i-1} + 106 f_{i-2} - 19 f_{i-3} \right]

Predictor-Corrector Method

Combine an explicit and an implicit method, since implicit ones rarely solve for wi+1w_{i+1} in closed form.

  • Predict
    Compute wi+1(p)w_{i+1}^{(p)} with an explicit method.
  • Correct
    Substitute wi+1(p)w_{i+1}^{(p)} into the right side of an implicit method to get wi+1w_{i+1}.

Standard choice: Runge-Kutta for starting values, Adams-Bashforth 4-step as predictor, Adams-Moulton 3-step as corrector.

Worked Examples

Example 1

Apply Adams-Bashforth 2-step to y′(t)=y−t2+1y'(t) = y - t^2 + 1, 0≤t≤20 \leq t \leq 2, y(0)=0.5y(0) = 0.5, with h=0.2h = 0.2. Exact solution y(t)=(t+1)2−0.5ety(t) = (t+1)^2 - 0.5 e^t.

Solution Reveal Hide

w0=0.5w_0 = 0.5 is given. Get w1w_1 with 1 step of order 4 Runge-Kutta.

k1=0.2 f(0, 0.5)=0.3k2=0.2 f(0.1, 0.5+0.15)=0.328k3=0.2 f(0.1, 0.5+0.164)=0.3308k4=0.2 f(0.2, 0.5+0.3308)=0.35816w1=0.5+16(k1+2k2+2k3+k4)=0.829293\begin{aligned} k_1 &= 0.2\, f(0,\ 0.5) = 0.3 \\ k_2 &= 0.2\, f(0.1,\ 0.5 + 0.15) = 0.328 \\ k_3 &= 0.2\, f(0.1,\ 0.5 + 0.164) = 0.3308 \\ k_4 &= 0.2\, f(0.2,\ 0.5 + 0.3308) = 0.35816 \\ w_1 &= 0.5 + \tfrac{1}{6}\left( k_1 + 2k_2 + 2k_3 + k_4 \right) = 0.829293 \end{aligned}

Step to t2=0.4t_2 = 0.4 by Adams-Bashforth

f0=w0−0.02+1=1.5f1=w1−0.22+1=1.789293w2=w1+0.22[3f1−f0]=0.829293+0.1 (3.867879)=1.216081\begin{aligned} f_0 &= w_0 - 0.0^2 + 1 = 1.5 \\ f_1 &= w_1 - 0.2^2 + 1 = 1.789293 \\ w_2 &= w_1 + \frac{0.2}{2}\left[ 3 f_1 - f_0 \right] = 0.829293 + 0.1\,(3.867879) = 1.216081 \end{aligned}

Exact y(0.4)≈1.214088y(0.4) \approx 1.214088.

Example 2

Use the Adams-Bashforth 2-step predictor and the Adams-Moulton 2-step corrector to find w2w_2 for y′=yy' = y, y(0)=1y(0) = 1, h=0.1h = 0.1, with w1=1.105171w_1 = 1.105171.

Solution Reveal Hide
f0=1,f1=1.105171f_0 = 1, \quad f_1 = 1.105171

Predict

w2(p)=w1+0.12[3(1.105171)−1]=1.220947w_2^{(p)} = w_1 + \frac{0.1}{2}\left[3(1.105171) - 1\right] = 1.220947

Correct

w2=w1+0.112[5(1.220947)+8(1.105171)−1]=1.221389w_2 = w_1 + \frac{0.1}{12}\left[5(1.220947) + 8(1.105171) - 1\right] = 1.221389

The exact value is e0.2=1.221403e^{0.2} = 1.221403. The corrector reduces the error from 4.6×10−44.6 \times 10^{-4} to 1.4×10−51.4 \times 10^{-5}.

Wolfram Mathematica notebooks

A companion notebook for this note. Open it in Mathematica and evaluate the cells to reproduce the worked example.

The summary notebook defines every function used in this module, with documentation and no examples, as a starting point for your own problems.

Written by November 3, 2026 2 min read
Was this helpful?