Gershgorin Circle Calculator

Introduction to Gershgorin circles for eigenvalue bounds

Gershgorin circles are one of the quickest ways to answer a practical linear algebra question: where could the eigenvalues of this matrix possibly be? Instead of solving the full characteristic polynomial, the theorem converts each row of a matrix into a disc in the complex plane. The diagonal entry becomes the disc’s center on the real axis, and the combined size of the off-diagonal entries becomes the radius. Once those discs are drawn, every eigenvalue must live inside their union. That makes Gershgorin’s theorem a fast screening tool for numerical methods, stability checks, matrix intuition, and classroom work where you want a reliable bound before doing heavier computation.

This calculator focuses on a 3×3 matrix because that size is common in teaching, quick hand verification, and small modeling problems. You enter the nine entries of the matrix, the calculator computes the three row radii, and the result reports one disc per row. If a diagonal entry is large and the off-diagonal couplings in that row are small, the corresponding disc is tight and informative. If the off-diagonal entries are large, the disc expands, which signals more uncertainty about where the associated eigenvalues may sit. In other words, the output tells you how strongly each row pulls eigenvalues away from its diagonal term.

Because the theorem is geometric, the result is especially useful for developing intuition. You are not merely seeing a number; you are seeing a center-and-radius description of how the matrix behaves. That is why this page explains the matrix inputs, the exact row formula, a worked example, and the limits of the method. Used correctly, Gershgorin discs are not a replacement for exact eigensolvers, but they are an excellent first estimate and a trustworthy bound.

What problem does this Gershgorin circle calculator solve for a matrix?

This Gershgorin circle calculator solves the problem of finding guaranteed eigenvalue regions without solving for the eigenvalues explicitly. For a general matrix, exact eigenvalue computation can involve a nontrivial polynomial or a numerical routine. Gershgorin’s theorem bypasses that by giving a safe enclosure. If you only need to know whether eigenvalues stay near the diagonal, whether they may cross into an unstable region, or whether row coupling is mild or severe, these discs can answer that question immediately.

In numerical linear algebra, this matters more often than it first appears. Before running an iterative method, you may want a rough sense of spectral spread. In a stability discussion, you may want to know whether all eigenvalues are plausibly clustered near positive diagonal entries. In a classroom setting, you may want to compare a nearly diagonal matrix with a strongly coupled one and see how the discs widen. The calculator gives that structural view in seconds. It is not about exact roots first; it is about turning a matrix into a geometric bound you can inspect and reason about.

How to use the Gershgorin circle calculator on a 3×3 matrix

This Gershgorin circle calculator is easiest to use if you think row by row. Enter the matrix entries exactly as they appear in your 3×3 matrix, keeping signs intact. The first row is a11, a12, a13, the second row is a21, a22, a23, and the third row is a31, a32, a33. Then click Compute Discs. The result area will list three discs, one for each row, with the center equal to the diagonal term and the radius equal to the sum of the absolute values of the other two entries in that row.

1. Type the first row of your matrix into a11, a12, and a13.
2. Type the second row into a21, a22, and a23.
3. Type the third row into a31, a32, and a33.
4. Submit the form to compute the three Gershgorin discs.
5. Read each line as center = diagonal entry and radius = off-diagonal magnitude sum.
6. Interpret the union of those discs as a guaranteed region containing all eigenvalues.

If you are checking your own hand work, compare each displayed radius with the corresponding row sum of absolute off-diagonal values. That quick check catches most input mistakes. If a sign error slips into an off-diagonal entry, the center will not change, but the radius still uses the absolute value, so you can verify whether your row sum looks reasonable.

Inputs: reading the matrix entries row by row

The inputs on this page are the entries of a real 3×3 matrix. The diagonal entries a11, a22, and a33 matter in a special way because they become the disc centers. If those diagonal values are far apart, you may get well-separated discs. If they are close together, even modest radii can create overlap. The off-diagonal entries measure coupling between coordinates. Their absolute values determine how much each disc expands away from its center.

A common beginner mistake is to think the sign of an off-diagonal entry changes the radius directly. It does not. In Gershgorin’s theorem, the radius uses absolute values, so both -3 and 3 contribute 3 to the radius. Another common mistake is to mix up rows and columns. This calculator uses the row version of the theorem, so disc 1 comes from row 1, disc 2 comes from row 2, and disc 3 comes from row 3. If you want the column version, you would instead sum the off-diagonal entries in each column, but that is not what this tool is computing.

When you choose inputs, try to notice the structure of your matrix before you calculate. A diagonal matrix gives zero radii, so every eigenvalue is exactly at a diagonal entry. A nearly diagonal matrix gives small radii and therefore tight bounds. A matrix with large off-diagonal terms can produce large overlapping discs, which means the theorem still gives a guarantee but not a very sharp one. That pattern is valuable information by itself: the discs tell you whether your matrix is spectrally easy to localize or intrinsically more entangled.

Formulas: the row-disc bounds behind the calculator

The row-disc formula used by this Gershgorin circle calculator is simple but powerful. For each row, first compute the radius by adding the absolute values of the off-diagonal entries in that row. Then place a disc in the complex plane centered at the diagonal entry. For a 3×3 matrix, the three radii are:

Once those radii are known, the discs are centered at a11, a22, and a33. For example, the first disc is:

The key theorem says that every eigenvalue of the matrix lies in at least one of the discs. At a higher level, you can still think of the calculator as taking several inputs and converting them through a defined rule into a result. The general idea behind many calculators is preserved below, and in this case the function f is specifically the Gershgorin row-sum mapping from matrix entries to disc centers and radii:

In the Gershgorin setting, the weighted-sum idea appears in each radius: the off-diagonal terms accumulate to measure how far an eigenvalue can stray from the diagonal center associated with that row.

Worked example: bounding eigenvalues for a sample 3×3 matrix

This worked example uses a concrete 3×3 matrix so the Gershgorin process feels less abstract. Suppose

A = [[4, -1, 0.5], [2, 3, -1], [0, 1, 5]]

For row 1, the center is a11 = 4 and the radius is |-1| + |0.5| = 1.5. So disc 1 is centered at 4 with radius 1.5. For row 2, the center is a22 = 3 and the radius is |2| + |-1| = 3. For row 3, the center is a33 = 5 and the radius is |0| + |1| = 1. The calculator would therefore report three discs: (4, 1.5), (3, 3), and (5, 1).

What does that mean geometrically? Along the real axis, disc 1 spans roughly from 2.5 to 5.5, disc 2 spans from 0 to 6, and disc 3 spans from 4 to 6 if you only look at the horizontal intercepts. The actual discs live in the complex plane, not just on the real line, so an eigenvalue could also have an imaginary part as long as it stays inside at least one disc. Even so, the interval view gives a quick sanity check: the eigenvalues cannot wander arbitrarily far from the diagonal values 4, 3, and 5 because the row sums restrict them.

This is also a good example of why overlap matters. Disc 2 is wide enough to overlap the other two discs, which means the theorem gives a broad guaranteed region rather than three isolated pockets. That does not make the result wrong; it simply means the matrix has enough row coupling that the simple bound is less sharp than it would be for a nearly diagonal matrix.

Comparison table: how one off-diagonal entry widens a disc

This comparison focuses on a specifically Gershgorin idea: changing a single off-diagonal term changes only the radius of its row disc, not the center. That makes sensitivity easy to see. In the table below, keep the sample matrix the same except for a23. Because disc 2 uses row 2, increasing the magnitude of a23 directly expands the radius of disc 2 by the same amount.

The important lesson is that diagonal entries place the centers, while off-diagonal magnitudes control how much uncertainty each row contributes. If a matrix model becomes more strongly coupled, you should expect wider Gershgorin discs.

How to interpret the Gershgorin disc result

The result panel reports each disc as a center and a radius. Read that output literally. A line such as Disc 1: center 4.0000, radius 1.5000 means every eigenvalue associated with the theorem’s guarantee must lie inside the disc centered at 4 in the complex plane with radius 1.5, or inside one of the other discs. The theorem is about the union of all reported discs. An eigenvalue does not need to belong to every disc; it only needs to belong to at least one of them.

Smaller radii usually mean stronger localization. If all three discs are narrow and well separated, you have a strong visual hint that eigenvalues cluster near the diagonal entries. If one disc is enormous and overlaps everything, the theorem is still valid but less informative. That is the right way to interpret a wide radius: not as a calculation problem, but as a structural statement that the row’s off-diagonal terms allow larger spectral movement away from the center.

Because this tool prints numerical values rather than drawing the discs directly, it is helpful to imagine the real-axis interval from center − radius to center + radius for each row. That interval is only a shadow of the full disc, but it often helps you compare rows quickly before doing a more detailed plot elsewhere.

Limitations of Gershgorin circle estimates

Gershgorin circle estimates are guaranteed bounds, but they are not exact eigenvalues and they are not always tight. That is the first limitation to keep in mind. A matrix can have eigenvalues in a small cluster even when one Gershgorin disc is quite large. The theorem only says where eigenvalues must be allowed to exist, not how tightly they actually concentrate inside the discs. In some problems, especially when off-diagonal terms are large, the guarantee can be noticeably conservative.

A second limitation is that this page uses the row version of the theorem. There is also a column version, and in some cases the column discs may be tighter or simply more convenient for the question you are asking. This calculator does not combine row and column information, nor does it apply more advanced refinements such as similarity scaling or block Gershgorin techniques. If you need sharper localization, those tools may help.

A third limitation is numerical interpretation. The calculator accepts real entries and returns the disc data, but it does not solve the characteristic equation, draw exact plots, or count how many eigenvalues fall into each connected component. The classical theorem does allow additional conclusions when discs are disjoint, but this page deliberately stays with the basic bound so the result remains transparent and fast.

Those limitations do not make the calculator weak. They define its proper role. Use it when you need a reliable first enclosure, a row-by-row diagnostic, or a teaching aid for understanding how diagonal dominance and off-diagonal coupling affect eigenvalue location. If you need exact eigenvalues, multiplicities, or high-precision spectral analysis, treat the discs as the starting geometry and then move on to a dedicated eigensolver.