Search

A recent paper by Steihaug and Wolfbrandt [6] gives a second-order method for stiff differential equations, which includes an error estimate. We show how, in most practical circumstances, the error estimate can be obtained more simply, and derive two third-order methods along similar lines. All the methods described are single-step and linearly implicit, and they are designed to avoid frequent evaluations of the Jacobian and inversion of the associated matrix.

A new method for the numerical solution of non linear parabolic equations is presented. The method is an extension of an existing algorithm for linear equations. Solutions are obtained in the form of a Chebyshev series, which is produced by approximating the partial differential equation by a set of ordinary differential equations over a small time interval. The method appears to be both accurate and economical.

A numerical method applied to the differential equation y′=λy can lead to a solution of the form , where z and w = λh satisfy an equation P(w, z) = 0. In general P(w,z) is a polynomial in both w and z, of degree M in w and N in z. Existing multistep and Runge-Kuta methods correspond to the cases M = 1 and N = 1 respectively. New methods are fottd by taking MS≧2N≧2. The approach here is first to find a suitable polynomial P(w,z) with the desired stability properties, and then to find a process which leads to this polynomial. Third- and fourth-order A- and L-stable processes are given of the semi-explicit arid linearly implicit types.

This paper solves the machine interference problem in which N automatic machines are maintained by a team of r operatives. Repair times are assumed to follow a negative exponential distribution, and running times for each of the machines are assumed to be independently and identically distributed. It is shown that the solution to this G/M/ r model is identical in most important respects to that for the M/M/ r model.

The matrix exponential exp(At) may be approximated to in some circumstances by the matrix function -p,q(At), where -p,q(u) is a Pade approximant to exp(u). The approximation breaks down if A has large latent roots, and in this case may be replaced by {-p,q(At/r)}r for some integer r. An error analysis of the latter approximation is given. A reply from the original author follows.