Insert an exact smooth solution into the linear multistep method and define its unscaled local truncation error by
Its Taylor expansion has coefficients
The method has order at least exactly when , since then for every sufficiently smooth . Necessity can be checked by inserting polynomials of successive degrees. This also corresponds to the usual truncation error .
On the other hand, expansion of the characteristic polynomials of a linear multistep method gives
Thus the exponential-symbol order criterion for a multistep method is
If “order ” is meant to be the exact rather than guaranteed order, one additionally requires . The first two vanishing conditions are the usual consistency of a numerical method relations , .
The relevant one of the characteristic polynomials of a linear multistep method is
The coefficients on the right are chosen to give third order, so consistency of a numerical method is already available. By the Dahlquist equivalence theorem, convergence with convergent starting values is equivalent to zero-stability: every root of must satisfy , and every root on the unit circle must be simple.
If , the quadratic is negative at and tends to as , so it has a root less than . If , its root product is , so at least one root has modulus greater than one. For , its real roots lie in . For , its conjugate roots have modulus .
At , the three roots are , with the unit-modulus roots simple. At , they are , again simple. The repeated quadratic root at is , strictly inside the unit circle, which the root condition for a multistep method permits. Therefore
This is the root stability of a cubic multistep polynomial; oscillatory parasitic roots at the endpoints do not invalidate convergence when their starting amplitudes tend to zero.

Articles by others on the same topic (0)

There are currently no matching articles.