When Greedy Algorithms are Perfect: the Matroid

Greedy algorithms are by far one of the easiest and most well-understood algorithmic techniques. There is a wealth of variations, but at its core the greedy algorithm optimizes something using the natural rule, “pick what looks best” at any step. So a greedy routing algorithm would say to a routing problem: “You want to visit all these locations with minimum travel time? Let’s start by going to the closest one. And from there to the next closest one. And so on.”

Because greedy algorithms are so simple, researchers have naturally made a big effort to understand their performance. Under what conditions will they actually solve the problem we’re trying to solve, or at least get close? In a previous post we gave some easy-to-state conditions under which greedy gives a good approximation, but the obvious question remains: can we characterize when greedy algorithms give an optimal solution to a problem?

The answer is yes, and the framework that enables us to do this is called a matroid. That is, if we can phrase the problem we’re trying to solve as a matroid, then the greedy algorithm is guaranteed to be optimal. Let’s start with an example when greedy is provably optimal: the minimum spanning tree problem. Throughout the article we’ll assume the reader is familiar with the very basics of linear algebra and graph theory (though we’ll remind ourselves what a minimum spanning tree is shortly). For a refresher, this blog has primers on both subjects. But first, some history.


Matroids were first introduced by Hassler Whitney in 1935, and independently discovered a little later by B.L. van der Waerden (a big name in combinatorics). They were both interested in devising a general description of “independence,” the properties of which are strikingly similar when specified in linear algebra and graph theory. Since then the study of matroids has blossomed into a large and beautiful theory, one part of which is the characterization of the greedy algorithm: greedy is optimal on a problem if and only if the problem can be represented as a matroid. Mathematicians have also characterized which matroids can be modeled as spanning trees of graphs (we will see this momentarily). As such, matroids have become a standard topic in the theory and practice of algorithms.

Minimum Spanning Trees

It is often natural in an undirected graph G = (V,E) to find a connected subset of edges that touch every vertex. As an example, if you’re working on a power network you might want to identify a “backbone” of the network so that you can use the backbone to cheaply travel from any node to any other node. Similarly, in a routing network (like the internet) it costs a lot of money to lay down cable, it’s in the interest of the internet service providers to design analogous backbones into their infrastructure.

A minimal subset of edges in a backbone like this is guaranteed to form a tree. This is simply because if you have a cycle in your subgraph then removing any edge on that cycle doesn’t break connectivity or the fact that you can get from any vertex to any other (and trees are the maximal subgraphs without cycles). As such, these “backbones” are called spanning trees. “Span” here means that you can get from any vertex to any other vertex, and it suggests the connection to linear algebra that we’ll describe later, and it’s a simple property of a tree that there is a unique path between any two vertices in the tree.

An example of a spanning tree

An example of a spanning tree

When your edges e \in E have nonnegative weights w_e \in \mathbb{R}^{\geq 0}, we can further ask to find a minimum cost spanning tree. The cost of a spanning tree T is just the sum of its edges, and it’s important enough of a definition to offset.

Definition: A minimum spanning tree T of a weighted graph G (with weights w_e \geq 0 for e \in E) is a spanning tree which minimizes the quantity

w(T) = \sum_{e \in T} w_e

There are a lot of algorithms to find minimal spanning trees, but one that will lead us to matroids is Kruskal’s algorithm. It’s quite simple. We’ll maintain a forest F in G, which is just a subgraph consisting of a bunch of trees that may or may not be connected. At the beginning F is just all the vertices with no edges. And then at each step we add to F the edge e whose weight is smallest and also does not introduce any cycles into F. If the input graph G is connected then this will always produce a minimal spanning tree.

Theorem: Kruskal’s algorithm produces a minimal spanning tree of a connected graph.

Proof. Call F_t the forest produced at step t of the algorithm. Then F_0 is the set of all vertices of G and F_{n-1} is the final forest output by Kruskal’s (as a quick exercise, prove all spanning trees on n vertices have n-1 edges, so we will stop after n-1 rounds). It’s clear that F_{n-1} is a tree because the algorithm guarantees no F_i will have a cycle. And any tree with n-1 edges is necessarily a spanning tree, because if some vertex were left out then there would be n-1 edges on a subgraph of n-1 vertices, necessarily causing a cycle somewhere in that subgraph.

Now we’ll prove that F_{n-1} has minimal cost. We’ll prove this in a similar manner to the general proof for matroids. Indeed, say you had a tree T whose cost is strictly less than that of F_{n-1} (we can also suppose that T is minimal, but this is not necessary). Pick the minimal weight edge e \in T that is not in F_{n-1}. Adding e to F_{n-1} introduces a unique cycle C in F_{n-1}. This cycle has some strange properties. First, e has the highest cost of any edge on C. For otherwise, Kruskal’s algorithm would have chosen it before the heavier weight edges. Second, there is another edge in C that’s not in T (because T was a tree it can’t have the entire cycle). Call such an edge e'. Now we can remove e' from F_{n-1} and add e. This can only increase the total cost of F_{n-1}, but this transformation produces a tree with one more edge in common with T than before. This contradicts that T had strictly lower weight than F_{n-1}, because repeating the process we described would eventually transform F_{n-1} into T exactly, while only increasing the total cost.


Just to recap, we defined sets of edges to be “good” if they did not contain a cycle, and a spanning tree is a maximal set of edges with this property. In this scenario, the greedy algorithm performed optimally at finding a spanning tree with minimal total cost.

Columns of Matrices

Now let’s consider a different kind of problem. Say I give you a matrix like this one:

\displaystyle A = \begin{pmatrix} 2 & 0 & 1 & -1 & 0 \\ 0 & -4 & 0 & 1 & 0 \\ 0 & 0 & 1 & 0 & 7 \end{pmatrix}

In the standard interpretation of linear algebra, this matrix represents a linear function f from one vector space V to another W, with the basis (v_1, \dots, v_5) of V being represented by columns and the basis (w_1, w_2, w_3) of W being represented by the rows. Column j tells you how to write f(v_j) as a linear combination of the w_i, and in so doing uniquely defines f.

Now one thing we want to calculate is the rank of this matrix. That is, what is the dimension of the image of V under f? By linear algebraic arguments we know that this is equivalent to asking “how many linearly independent columns of A can we find”? An interesting consequence is that if you have two sets of columns that are both linearly independent and maximally so (adding any other column to either set would necessarily introduce a dependence in that set), then these two sets have the same size. This is part of why the rank of a matrix is well-defined.

If we were to give the columns of A costs, then we could ask about finding the minimal-cost maximally-independent column set. It sounds like a mouthful, but it’s exactly the same idea as with spanning trees: we want a set of vectors that spans the whole column space of A, but contains no “cycles” (linearly dependent combinations), and we want the cheapest such set.

So we have two kinds of “independence systems” that seem to be related. One interesting question we can ask is whether these kinds of independence systems are “the same” in a reasonable way. Hardcore readers of this blog may see the connection quite quickly. For any graph G = (V,E), there is a natural linear map from E to V, so that a linear dependence among the columns (edges) corresponds to a cycle in G. This map is called the incidence matrix by combinatorialists and the first boundary map by topologists.

The map is easy to construct: for each edge e = (v_i,v_j) you add a column with a 1 in the j-th row and a -1 in the i-th row. Then taking a sum of edges gives you zero if and only if the edges form a cycle. So we can think of a set of edges as “independent” if they don’t contain a cycle. It’s a little bit less general than independence over \mathbb{R}, but you can make it exactly the same kind of independence if you change your field from real numbers to \mathbb{Z}/2\mathbb{Z}. We won’t do this because it will detract from our end goal (to analyze greedy algorithms in realistic settings), but for further reading this survey of Oxley assumes that perspective.

So with the recognition of how similar these notions of independence are, we are ready to define matroids.

The Matroid

So far we’ve seen two kinds of independence: “sets of edges with no cycles” (also called forests) and “sets of linearly independent vectors.” Both of these share two trivial properties: there are always nonempty independent sets, and every subset of an independent set is independent. We will call any family of subsets with this property an independence system.

Definition: Let X be a finite set. An independence system over X is a family \mathscr{I} of subsets of X with the following two properties.

  1. \mathscr{I} is nonempty.
  2. If I \in \mathscr{I}, then so is every subset of I.

This is too general to characterize greedy algorithms, so we need one more property shared by our examples. There are a few things we do, but here’s one nice property that turns out to be enough.

Definition: A matroid M = (X, \mathscr{I}) is a set X and an independence system \mathscr{I} over X with the following property:

If A, B are in \mathscr{I} with |A| = |B| + 1, then there is an element x \in A \setminus B such that B \cup \{ a \} \in \mathscr{I}.

In other words, this property says if I have an independent set that is not maximally independent, I can grow the set by adding some suitably-chosen element from a larger independent set. We’ll call this the extension property. For a warmup exercise, let’s prove that the extension property is equivalent to the following (assuming the other properties of a matroid):

For every subset Y \subset X, all maximal independent sets contained in Y have equal size.

Proof. For one direction, if you have two maximal sets A, B \subset Y \subset X that are not the same size (say A is bigger), then you can take any subset of A whose size is exactly |B| + 1, and use the extension property to make B larger, a contradiction. For the other direction, say that I know all maximal independent sets of any Y \subset X have the same size, and you give me A, B \subset X. I need to find an a \in A \setminus B that I can add to B and keep it independent. What I do is take the subset Y = A \cup B. Now the sizes of A, B don’t change, but B can’t be maximal inside Y because it’s smaller than A (A might not be maximal either, but it’s still independent). And the only way to extend B is by adding something from A, as desired.


So we can use the extension property and the cardinality property interchangeably when talking about matroids. Continuing to connect matroid language to linear algebra and graph theory, the maximal independent sets of a matroid are called bases, the size of any basis is the rank of the matroid, and the minimal dependent sets are called circuits. In fact, you can characterize matroids in terms of the properties of their circuits, which are dual to the properties of bases (and hence all independent sets) in a very concrete sense.

But while you could spend all day characterizing the many kinds of matroids and comatroids out there, we are still faced with the task of seeing how the greedy algorithm performs on a matroid. That is, suppose that your matroid M = (X, \mathscr{I}) has a nonnegative real number w(x) associated with each x \in X. And suppose we had a black-box function to determine if a given set S \subset X is independent. Then the greedy algorithm maintains a set B, and at every step adds a minimum weight element that maintains the independence of B. If we measure the cost of a subset by the sum of the weights of its elements, then the question is whether the greedy algorithm finds a minimum weight basis of the matroid.

The answer is even better than yes. In fact, the answer is that the greedy algorithm performs perfectly if and only if the problem is a matroid! More rigorously,

Theorem: Suppose that M = (X, \mathscr{I}) is an independence system, and that we have a black-box algorithm to determine whether a given set is independent. Define the greedy algorithm to iteratively adds the cheapest element of X that maintains independence. Then the greedy algorithm produces a maximally independent set S of minimal cost for every nonnegative cost function on X, if and only if M is a matroid.

It’s clear that the algorithm will produce a set that is maximally independent. The only question is whether what it produces has minimum weight among all maximally independent sets. We’ll break the theorem into the two directions of the “if and only if”:

Part 1: If M is a matroid, then greedy works perfectly no matter the cost function.
Part 2: If greedy works perfectly for every cost function, then M is a matroid.

Proof of Part 1.

Call the cost function w : X \to \mathbb{R}^{\geq 0}, and suppose that the greedy algorithm picks elements B = \{ x_1, x_2, \dots, x_r \} (in that order). It’s easy to see that w(x_1) \leq w(x_2) \leq \dots \leq w(x_r). Now if you give me any list of r independent elements y_1, y_2, \dots, y_r \in X that has w(y_1) \leq \dots \leq w(y_r), I claim that w(x_i) \leq w(y_i) for all i. This proves what we want, because if there were a basis of size r with smaller weight, sorting its elements by weight would give a list contradicting this claim.

To prove the claim, suppose to the contrary that it were false, and for some k we have w(x_k) > w(y_k). Moreover, pick the smallest k for which this is true. Note k > 1, and so we can look at the special sets S = \{ x_1, \dots, x_{k-1} \} and T = \{ y_1, \dots, y_k \}. Now |T| = |S|+1, so by the matroid property there is some j between 1 and r so that S \cup \{ y_j \} is an independent set (and y_j is not in S). But then w(y_j) \leq w(y_k) < w(x_k), and so the greedy algorithm would have picked y_j before it picks x_k (and the strict inequality means they’re different elements). This contradicts how the greedy algorithm runs, and hence proves the claim.

Proof of Part 2.

We’ll prove this contrapositively as follows. Suppose we have our independence system and it doesn’t satisfy the last matroid condition. Then we’ll construct a special weight function that causes the greedy algorithm to fail. So let A,B be independent sets with |A| = |B| + 1, but for every a \in A \setminus B adding a to B never gives you an independent set.

Now what we’ll do is define our weight function so that the greedy algorithm picks the elements we want in the order we want (roughly). In particular, we’ll assign all elements of A \cap B a tiny weight we’ll call w_1. For elements of B - A we’ll use w_2, and for A - B we’ll use w_3, with w_4 for everything else. In a more compact notation:


We need two things for this weight function to screw up the greedy algorithm. The first is that w_1 < w_2 < w_3 < w_4, so that greedy picks the elements in the order we want. Note that this means it’ll first pick all of A \cap B, and then all of B - A, and by assumption it won’t be able to pick anything from A - B, but since B is assumed to be non-maximal, we have to pick at least one element from X - (A \cup B) and pay w_4 for it.

So the second thing we want is that the cost of doing greedy is worse than picking any maximally independent set that contains A (and we know that there has to be some maximal independent set containing A). In other words, if we call m the size of a maximally independent set, we want

\displaystyle |A \cap B| w_1 + |B-A|w_2 + (m - |B|)w_4 > |A \cap B|w_1 + |A-B|w_3 + (m-|A|)w_4

This can be rearranged (using the fact that |A| = |B|+1) to

\displaystyle w_4 > |A-B|w_3 - |B-A|w_2

The point here is that the greedy picks too many elements of weight w_4, since if we were to start by taking all of A (instead of all of B), then we could get by with one fewer. That might not be optimal, but it’s better than greedy and that’s enough for the proof.

So we just need to make w_4 large enough to make this inequality hold, while still maintaining w_2 < w_3. There are probably many ways to do this, and here’s one. Pick some 0 < \varepsilon < 1, and set


It’s trivial that w_1 < w_2 and w_3 < w_4. For the rest we need some observations. First, the fact that |A-B| = |B-A| + 1 implies that w_2 < w_3. Second, both |A-B| and |B-A| are nonempty, since otherwise the second property of independence systems would contradict our assumption that augmenting B with elements of A breaks independence. Using this, we can divide by these quantities to get

\displaystyle w_4 = 2 > 1 = \frac{|A-B|(1 + \varepsilon)}{|A-B|} - \frac{|B-A|\varepsilon}{|B-A|}

This proves the claim and finishes the proof.


As a side note, we proved everything here with respect to minimizing the sum of the weights, but one can prove an identical theorem for maximization. The only part that’s really different is picking the clever weight function in part 2. In fact, you can convert between the two by defining a new weight function that subtracts the old weights from some fixed number N that is larger than any of the original weights. So these two problems really are the same thing.

This is pretty amazing! So if you can prove your problem is a matroid then you have an awesome algorithm automatically. And if you run the greedy algorithm for fun and it seems like it works all the time, then that may be hinting that your problem is a matroid. This is one of the best situations one could possibly hope for.

But as usual, there are a few caveats to consider. They are both related to efficiency. The first is the black box algorithm for determining if a set is independent. In a problem like minimum spanning tree or finding independent columns of a matrix, there are polynomial time algorithms for determining independence. These two can both be done, for example, with Gaussian elimination. But there’s nothing to stop our favorite matroid from requiring an exponential amount of time to check if a set is independent. This makes greedy all but useless, since we need to check for independence many times in every round.

Another, perhaps subtler, issue is that the size of the ground set X might be exponentially larger than the rank of the matroid. In other words, at every step our greedy algorithm needs to find a new element to add to the set it’s building up. But there could be such a huge ocean of candidates, all but a few of which break independence. In practice an algorithm might be working with X implicitly, so we could still hope to solve the problem if we had enough knowledge to speed up the search for a new element.

There are still other concerns. For example, a naive approach to implementing greedy takes quadratic time, since you may have to look through every element of X to find the minimum-cost guy to add. What if you just have to have faster runtime than O(n^2)? You can still be interested in finding more efficient algorithms that still perform perfectly, and to the best of my knowledge there’s nothing that says that greedy is the only exact algorithm for your favorite matroid. And then there are models where you don’t have direct/random access to the input, and lots of other ways that you can improve on greedy. But those stories are for another time.

Until then!

Linear Algebra – A Primer

Story Time

Linear algebra was founded around the same time as Calculus (think Leibniz, circa 1700) solely for the purpose of solving general systems of linear equations. The coefficients of a system were written in a grid form, with rows corresponding to equations and columns to the unknown variables. Using a computational tool called the determinant (an awkward, but computable formula involving only the coefficients of the equations in a system), researchers were able to solve these systems, opening a world of information about the positions of celestial bodies and large-scale measurements (of geodesic arcs) on the surface of the earth.

By the 1850’s, Arthur Cayley was representing matrices as abstract objects. He defined matrix multiplication and nurtured matrix theory as its own field, recognizing a vast wealth of theoretical knowledge underlying the theory of determinants. Around turn of the century, a formal system of vector algebra was invented which relied heavily on interpreting matrices as so-called linear transformations. Linear transformations are intuitively those maps of everyday space (\mathbb{R}^n) which preserve “linear” things. Specifically, they send lines to lines, planes to planes, etc., and they preserve the origin (one which does not preserve the origin is very similar but has a different name; see Affine Transformation). Soon enough the mathematical focus shifted to the foundations of such an algebra, and later with the advent of computers to rapid calculations in one.


Linear algebra sits at the crossroads of many areas of mathematics. Keeping close to its roots, linear algebra is primarily a tool for computation. Unsurprisingly, a huge chunk of mathematical research has been solely to phrase things in terms of matrices and their associated linear transformations. For instance, an undirected graph on n vertices can be modeled as a matrix of integer entries, with the i,j entry containing the number of edges from vertex i to vertex j. This is called the adjacency matrix of a graph. Suddenly, a wealth of information about the graph translates to simple matrix computations. For instance, we can compute the number of paths from one vertex to another of length m as the appropriate entry of A^m. (more formally,these are walks, which are allowed to repeat edge traversals and visited vertices)

Even in advanced, purely theoretical mathematics, objects are commonly represented in terms of coordinates in some vector space, and are subsequently studied using all of the great things we know about linear transformations and their matrices. And so, without further ado, we will present the terminology and working concepts necessary for the content elsewhere in this blog.

Vector Spaces

The setting for all of linear algebra is in some vector space. Intuitively this is just a collection of objects, which we call vectors, with some rules on how you can combine vectors to get other vectors. This treatment wouldn’t do that idea justice without an axiomatic definition, so here it is.

Definition: A vector space is a quadruple (V, F, +, \cdot), where V is a set of vectors (points in our space), F is a scalar field (coefficients), +:V \times V \to V is a commutative, associative operation to combine vectors, and \cdot: F \times V \to V is an operation to “scale” vectors. In addition, we need the following properties to hold:

  • Addition and multiplication distribute (as we are used to with traditional algebra).
  • There must be an additive identity, which we call 0, giving 0 + v = v for all v \in V.
  • Every vector must have an additive inverse (every v has some w with v + w = 0).

This is a lot to swallow at first, but it is general for a good reason: there are tons of different kinds of vector spaces! Many of these are surprising and counter-intuitive. For our purposes, however, we may stick with the nice, small vector spaces. So here is a simplified definition that will suffice:

Definition: vector space is a set V of vectors which are fixed-length lists of real numbers (v_1, v_2, \dots , v_n) \in \mathbb{R}^n, where addition between vectors is componentwise, we may scale vectors by any real number, and the following properties hold:

  • Addition and multiplication distribute (as above).
  • (0,0,0, \dots, 0) is the additive identity.
  • (-v_1, -v_2, \dots , -v_n) is the unique additive inverse of (v_1, v_2, \dots , v_n).

Hopefully this is much more familiar to what we think of as “vectors,” and with the understanding that we are viewing it as a vector space, we just call it \mathbb{R}^n. The closure of operations gives us a nice way to characterize “any combination” of vectors in a vector space.

Definition: A linear combination of vectors in a vector space V is the vector

a_1v_1 + a_2v_2 + \dots + a_kv_k

for some positive integer k, scalars a_i, and vectors v_i.

We may speak of the span of a set of vectors as the set of all possible linear combinations of those vectors. Furthermore, we call a set of vectors linearly independent if no vector in the list is in the span of the others. For example, (1,0,0), (0,1,0), and (0,0,1) are linearly independent in \mathbb{R}^3. Specifically, (1,0,0) cannot be written as a(0,1,0) + b(0,0,1) = (0,a,b) for any scalars a,b \in F, and the other two vectors are similarly so.

As usual, we may describe subspaces of a vector space, which are just subsets of V which are themselves vector spaces with the inherited operations. The simplest examples of these are lines, planes, and hyperplanes through the origin in \mathbb{R}^n. Consequently, we may identify \mathbb{R}^n as a subspace of \mathbb{R}^m for any n \leq m.

One of the first things we want to ask about a vector space is “how big is it?” While most instances of vector spaces we will see have uncountably many elements, we can characterize “size” in terms of a different metric: the size of a basis.

Definition: A list of vectors (v_1, v_2, \dots v_n) is a basis for V if its elements are linearly independent, and their span is V. The dimension of a vector space is the length of any basis.

For \mathbb{R}^n, and similarly all finite-dimensional vector spaces, it is easy to prove that all bases have the same length, and hence dimension is well-defined. Further, \mathbb{R}^n admits a very natural basis, often called the standard basis:

e_1 = (1,0, \dots, 0)
e_2 = (0,1, \dots, 0)
e_n = (0,0, \dots, 1)

These are best visualized as the coordinate axes in \mathbb{R}^n, and it strokes our intuition as to what a basis should be, because any vector in \mathbb{R}^n can be broken down uniquely into a sum of scalar multiples of these unit coordinates. Indeed, this is true of any basis (due to linear independence). Given a fixed basis for V, every vector v \in V may be uniquely written as a linear combination of basis vectors.

Linear Transformations and their Matrix Representations

Moving quickly toward the heart of linear algebra, we may speak of linear transformations (interchangeably, linear maps) between two vector spaces:

Definition: A function f : V \to W is a linear map if it preserves the operations of addition and scalar multiplication. In other words, for all v, w \in V, c \in F, f(v+w) = f(v)+f(w) and f(cv) = cf(v).

Examples are bountiful; some geometrically inspired ones include rotations about the origin, shears, and scalings. These are functions you’d likely see in an image manipulation program like photoshop. From this we can prove a few basic facts, like that every linear map sends 0 to 0 and additive inverses to additive inverses (try it as an exercise).

One remarkable fact that helps us characterize linear maps is that every linear map is determined completely by what it does to a basis. Since every vector x \in V is a linear combination of basis elements, say x=a_1v_1 + \dots + a_nv_n, we see that a linear map plays nicely:

f(x) = f(a_1v_1 + \dots + a_nv_n) = a_1f(v_1) + \dots + a_nf(v_n)

In other words, if we know what f does to a basis, then we know everything about f. In order to aid our computations, we write what f does to each basis vector in a tabular form. To elaborate on the vague word “does,” we need to also fix a basis of our target vector space W, say (w_1, \dots , w_m), and describe each f(v_i) in terms of this basis. We write it in tabular form, as follows:

\begin{pmatrix} | & | & \mathbf{ } & | \\ f(v_1) & f(v_2) & \dots & f(v_n) \\ | & | & \mathbf{ } & | \end{pmatrix}

The jth column corresponds to f(v_j), and the ith row corresponds to the ith coefficient in the expansion of f(v_j) in terms of the basis for W. Here the vertical bars indicate that each element is a column of scalars. We will do an extended example to make this clear.

Consider the map f on \mathbb{R}^3 defined as (x,y,z) \mapsto (y,x,2z+y). It is easy to check this map is linear, and using the standard basis we see that

f(1,0,0) = (0,1,0),
f(0,1,0) = (1,0,1), and
f(0,0,1) = (0,0,2).


f(e_1) = e_2f(e_2) = e_1 + e_3, and f(e_3) = 2e_3.

Hence, the matrix representation of f with respect to the standard basis is

A = \begin{pmatrix} 0 & 1 & 0 \\ 1 & 0 & 0 \\ 0 & 1 & 2 \end{pmatrix}

Now we see that if we take a (column) vector x, and multiply it on the left by our matrix A, the resulting vector is precisely the coordinate representation of f(x) with respect to the basis for W. In fact, the rules for matrix multiplication were constructed very particularly so that this would be the case. In this way, we may arbitrarily switch between viewing f as a transformation and a vector computation. Compositions of linear maps translate to multiplication of two matrices, and matrix inversion (if it exists) is precisely function inversion.

Of course, there are many different bases we could have chosen. Even though we are going from \mathbb{R}^3 \to \mathbb{R}^3, the column basis could be different from the row basis. Fortunately for our purposes, we are not going to consider what basis is appropriate to choose. All that matters is that fixing a basis, the matrix representation of a linear map is unique, and so we may interchange the notation freely. Even so, the truly interesting things about matrices are those properties which are true no matter which basis we prefer to use.

Eigenvectors and Eigenvalues

Definition: A scalar \lambda \in F is an eigenvalue for the linear map A if there exists a non-zero vector v \in V with Av = \lambda v. Any such vector v which satisfies this equation is said to be an eigenvector of A corresponding to \lambda.

Eigenvectors and eigenvalues have a huge number of applications, including facial recognition software, geology, quantum mechanics, and web search. So being able to find them quickly is of great significance to researchers and engineers. What’s interesting is that while eigenvectors depend on a choice of basis, eigenvalues do not. We prove this now:

Proposition: If A and B are different representations of the same linear map, then any eigenvalue of B is an eigenvalue of A.

Proof. It turns out that the process of “changing a basis” can be boiled down to matrix multiplication. Specifically, if A and B are two different matrix representations of the same linear map, we have the existence of some invertible matrix P such that A = PBP^{-1}, or AP = PB. As a result, if v is an eigenvector for B corresponding to the eigenvalue \lambda, then for some APv = PBv = P \lambda v = \lambda Pv and so A(Pv) = \lambda(Pv), and Pv is an eigenvector for A corresponding to \lambda as well. This proves that eigenvalues are invariant with respect to a change of basis, as desired. \square

The point of this is that we can choose whatever basis we want to work with, and compute the eigenvalues where we’re most comfortable. For instance, if we choose a basis that gives the following diagonal representation,

A = \begin{pmatrix} 1 & 0 & 0 \\ 0 & 2 & 0 \\ 0 & 0 & 3 \end{pmatrix}

then we can just eyeball that the eigenvalues are 1, 2, and 3. In fact, there are some very deep theorems in linear algebra that concern the existence and uniqueness of certain matrix representations. For a more in-depth treatment, see Axler, Linear Algebra Done Right. We will cover all the necessary information in the relevant posts, but until then, we are absolutely pooped from typing. Until next time!