Working Through Sauer's Numerical Analysis Problem Sets
The textbook by Mathews and Sauer is one of those books that shows up on every numerical analysis reading list. It covers the standard ground—root finding, interpolation, numerical integration, ODEs, linear systems—and the problems range from tedious to actually interesting depending on which chapter you are in. The solution manual is useful, but it has some quirks you need to know about before you rely on it for anything serious. I ran into a specific issue last semester when a student was working through Chapter 5, the section on Newton-Raphson and its variants. Problem 5.37 asks you to implement a modified Newton method where the Jacobian is frozen at the initial guess and reused for several iterations. The solution manual gives a final answer rounded to six decimal places, but it skips the step where the iteration stalls because the approximate Jacobian becomes nearly singular after four iterations. You can't tell that from the back-of-the-book answer alone. I had to write out the full trace and plot the residual norm to show where the breakdown happened. That kind of gap is the most frustrating part of this manual—it assumes you will catch things on your own. Another thing nobody warns you about: some editions have inconsistencies between problem numbers. Edition 2 and edition 3 moved around the boundary value problem set, and a few solutions reference equations from a different chapter than the one the problem appears in. If you are cross-referencing and the solution looks like it belongs to a different problem, check your edition first before you assume the manual is wrong. I spent an afternoon chasing a phantom error in the Runge-Kutta section before I realized the problem number had shifted by two.
The manual is strongest on the introductory chapters. Root finding, fixed point iteration, basic interpolation—these are straightforward and the solutions are generally correct with reasonable detail. The linear algebra section is decent too, though you will notice the authors prefer certain algorithmic paths over others. They tend to favor Gaussian elimination with partial pivoting as the default and rarely show alternatives like Householder QR unless the problem specifically calls for it. That means if your course is emphasizing QR factorization or iterative methods, you will not find parallel coverage in the manual. On the ODE side, the solutions assume you are comfortable translating pseudo-code into actual code. The manual gives algorithm steps, not full programs. I once had someone complain that the Euler method solution for problem 8.14 did not match their Python output, and the discrepancy turned out to be a simple indexing mistake—they were evaluating the derivative at n plus one instead of at n. The manual itself was fine. This is the general pattern: the solutions are correct, but they are written for people who are already doing the work, not for people who need the work done for them. If you are using this manually alongside the textbook, here is a practical approach. Work the problem first without looking at the solution. Write out the algorithm on paper before you touch any code. Then check your work against the manual. When the manual's answer does not match yours, do not immediately assume you are wrong. Recalculate by hand for small cases—say, a two-step Runge-Kutta or a three-point Newton iterate—so you can see exactly where your logic diverges from the printed solution. That process usually takes you ten to fifteen minutes and teaches you more than blindly accepting the answer ever would.
There are also legitimate gaps in coverage. The manual barely touches on adaptive quadrature, spectral methods, or preconditioned Krylov solvers. If your syllabus goes into any of those areas, the solution manual will not help you much. In those cases, pairing it with online resources or lecture notes is necessary. The book's problem sets themselves are worth doing even without the manual, especially the computational projects at the end of each chapter. Those are where you actually learn something. Availability varies by region and retailer. Some universities license the instructor's copy and provide it through their course platform. Others leave it to students to source independently. The copyright is held by Pearson, and unauthorized distribution is not something I am going to link to here. If you need the manual, the safest route is through your institution or an authorized academic reseller. The bottom line is that the Sauer solution manual is a supplementary tool, not a replacement for working through the material. It is accurate for what it covers, incomplete for what it omits, and occasionally misleading about which edition its answers apply to. Use it wisely and you will save yourself a lot of time. Use it carelessly and you will walk into exams unprepared for the problems the textbook leaves deliberately unsolved.
Get the Full Details
