1.4 The Matrix Equation Ax = b We can write linear combinations as multiplication of a matrix with a vector. Definition: If A is an m x n matrix with columns a1, a2 ,, an and if x is in Rn, then the product of A and x denoted by Ax is a linear combination or the columns of A using the corresponding entries of x as weights. That is, x1 Ax a1 a2 an xn x1a1 x2a2 xn an For example, 1 4 1 4 3 2 7 7 3 6 2 6 0 5 0 5 7 24 31 21 - 12 9 0 - 30 30 Example: Consider: 2 3 4 9 3 1 0 2 We can write the system of equations corresponding to the augmented matrix: 2 x1 3x2 4 x3 9 3x1 x2 2 Then express the system of equations in vector form 2 3 4 9 x1 x2 x3 3 1 0 2 Then express this vector equation in the form Ax = b where b is a 2 x 1 vector. x1 2 3 4 9 3 1 0 x2 2 x 3 ►We now have three equivalent ways of viewing a linear system: 1. as a system of linear equations 2. as a vector equation 3. as a matrix equation Theorem 3 If A is an m x n matrix with columns a1, a2 ,, an and if b is in Rm, then the matrix equation Ax = b has the same solution set as the vector equation is x1v1 x2 v 2 xn v n b which in turn has the same solution set as the system of linear equations whose augmented matrix is a1 a2 a3 b . Useful fact: The equation Ax = b has a solution if an only if b is a linear combination of the columns of A. Example: Let b1 1 4 5 A 3 11 14 b b2 b3 2 8 10 , Is Ax = b consistent for all b? Create the augmented matrix corresponding to Ax = b. 1 4 5 b1 3 11 14 b 2 2 8 10 b3 R2+3R1 R3 + (-2R1) 1 0 0 4 5 1 1 0 0 b2 3b1 b3 2b1 b1 Is this consistent for all b? There are values of b1 ,b3 for which b3 2b1 0 . So, Ax = b in inconsistent for some b. Note: b3 2b1 0 is a plane in R3. If any b in R3 can be expressed as linear combination of the columns of A, then we say the columns of A span R3. Definition: We say that the columns of a1 a2 a p span Rm if every vector b in Rm is a linear combination of a1 , a 2 , , a p , and Span{ a1 , a 2 , , a p }= Rm. Theorem 1.4 Let A be an m x n matrix. Then the following statements are logically equivalent. (That is, for a particular A, either they are all true or they are all false.) a) For each b in Rm, Ax = b has a solution. b) Each b in Rm is a linear combination of the columns of A. c) The columns of A span Rm. d) A has a pivot position in every row. Proof: c ↔ b: The columns of A span Rm if and only if each b in Rm is a linear combination of the columns of A, by the definition of span. b ↔ a: Each b in Rm is a linear combination of the columns of A if and only if the equation Ax = b has a solution for each b in Rm, by the Useful Fact. We have c ↔ b and b ↔ a so c ↔ a All we need to show is d ↔ a, b, or c. d ↔ a is easiest. First d → a: A has a pivot position in every row. Let U be a row echelon form of A. Given b in Rm, we can write the augmented matrix [A b] whose columns are the columns of A and the vector b. We can row reduce this matrix to [U d] for some d in Rm. [U d] is now an augmented matrix in row echelon form. A has a pivot position in every row, so U has a pivot in every row. Thus, there is no pivot in d (by Homework problem 1.2.25). Thus, by Th 1.2, the system represented by [A b] is consistent, and the equation Ax = b has a solution so (a) is true. a → d: it is easier to prove the contrapositive: If (d) is not true, then (a) is not true: that is, we need to prove that if there is a row in A that has no pivot position, there exists a vector b in Rm, for which the equation Ax = b has no solution. We need to find that b. Say (d) is false. Then there is a row in A that does not have a pivot position, so the last row in U is all zeros since U is a row echelon form of A. Let d be any vector with a 1 in its last entry. Then [U d] represents an inconsistent system since the far right column has a pivot. Recall: row operations are reversible, so [U d] can be transformed into [A b] for some b. The new system represented by [A b] or Ax = b is inconsistent, so (a) is false. Since not d → not a, a → d, and we are done. QED Example: Let b1 1 2 b b2 A 3 4 b3 5 6 and Is Ax = b consistent for all possible b? Th. 1.4 says the equation is consistent if A has a pivot in each row. But, there are three rows in A and only two columns, so A can have at most two pivots.Thus, the equation is inconsistent for all possible b. Example: Do the columns of 1 2 3 A 2 4 6 3 span R ? 0 3 9 Th. 1.4 says A has a pivot position in every row if and only if the columns of A span R3. Row 2 has no pivot since it is a multiple of Row 1. Or 1 0 - 3 rref A 0 1 3 0 0 0 And Row 3 has no pivot. Since there are only two rows with pivots, the columns of A do not spen R3. ►There is another rule for computing Ax. Read Examples 4 and 5 on p44-45 to see it. Theorem 1.5: If A is an m x n matrix and u and v are vectors in Rn, and c is a scalar, then a. A(u + v) = Au + Av; b. A(cu) = c(Au) Proof of a: Au v u1 v1 u v 2 2 a1 a2 an un vn u1 v1 a1 u2 v2 a2 un vn a n u1a1 v1a1 un an vn an u1a1 unan v1a1 vnan Au Av