Dr. Neal, WKU Linear Combinations in Rn MATH 307 n Let u1, u2 , . . ., um be a collection of m distinct vectors in R . Given another vector v in Rn , we wish to test whether or not v is a linear combination of u1, u2 , . . ., um . That is, do there exist scalars c1 , c2 , . . ., cm such that € € € € v = c1 u1 + c 2 u2 + . . . + c m um ? € € € € n Each vector ui has n coordinates in R , so we may write € € u1 = ( u11 , u12 , . . ., u1n ) u2 = ( u21 , u2 2 , . . ., u2 n ) . . . . . um = ( um1 , um 2 , . . ., um n ). € € Likewise we may write v = ( v1, v2 , . . ., vn ). € € In order for v to be a linear combination of u1, u2 , . . ., um , then for some scalars c1 , have c1 u1 + c 2 u2 + . . . + c m um = v . Thus, c2 , . . ., cm we must € € . . ., u2 n )€+ . . . + cm ( um1 , um 2 , . . ., um n ) c1 ( u11 , u12 , . . ., u1n ) + c2 ( u21 , u2€2 , € € € = ( v1, v2 , . . ., vn ). Adding componentwise we obtain c1 u11 + c2 u21 + . . . + c1 u12 + c2 u2 2 + . . . + . . c1 u1n + c2 u2 n + . . . + cm um1 = v1 cm um 2 = v2 cm um n = vn We thereby obtain a system of n equations and m unknowns. The m unknowns are the scalars c1 , c2 , . . ., cm . The associated augmented matrix n comes from the coordinates of the vectors in R where the vectors are written in column form: € u1 u2 . . . u11 u12 . . u 1n u21 u2 2 . . u2 n . . . . . . . . . . um v . um1 v1 . um 2 v 2 . . . . . . . um n v n Dr. Neal, WKU • If there is no solution to this system, then v is not a linear combination of the vectors u1, u2 , . . ., um . • If there is a solution to the system, then v is a linear combination of u1, u2 , . . ., um € € € m € we can find scalars c , c , . . ., c such that v = ∑ ci ui . and m 1 2 € i =1 3 Example 1. In R , let u1 = (1, –2, 4), u2€= (2, –4, 6), and vector v = (2, 4, 6) is a linear combination of u1, u2 , and € € € € u = (–1, 3, 8). Determine if the 3 u3. If so, then write it as such. Solution. We check to see if there is a solution to the associated system of equations: € € € € € € € u1 u2 u3 v 1 0 0 −88 1 2 −1 2 → rref → 0 1 0 49 −2 −4 3 4 0 0 1 8 4 6 8 6 Thus, there is a solution to the system and (2, 4, 6) is a linear combination of u1, u2 , and u3. We see that (2, 4, 6) = –88 (1, –2, 4) + 49 (2, –4, 6) + 8 (–1, 3, 8), or in simpler € notation, v = −88 u1 + 49 u2 + 8 u3 . € € € n Note: When we have n vectors in R , then we can write them as columns to form an n × n matrix of coefficients A . If det (A) ≠ 0, then every system of equations AX = B will € n have a unique solution. Thus, any other vector in R will be a linear combination of these first n vectors. 1 2 −1 In the preceding example, A = −2 −4 3 and det (A) = 2 ≠ 0. Thus every vector 4 6 8 3 in R is some linear combination of the three vectors u1 = (1, –2, 4), u2 = (2, –4, 6), and u3 = (–1, 3, 8). € € 4 Example 2. In R , let u1 = (1, –2, 4, 2), u2 = (2, –4, 6, 3), and u3 = (–1, 3, 8, 2). Determine if v = (2, 4, 6, 8) is a linear combination of u1, u2 , and u3. € € Solution. We check to see if there is a solution to the associated system of equations: € € € € 1 1 0 0 0 2 €−1 €2 −2 −4 3 4 0 1 0 0 → rref → 6 8 6 4 0 0 1 0 2 0 0 0 1 3 2 8 Dr. Neal, WKU no solution to the system, then The last row gives an inconsistency. Because there is v = (2, 4, 6, 8) is not a linear combination of u1, u2 , and u3. Because we must solve a system of equations in order to determine whether or not a vector is a linear combination of other vectors, we may not always have a unique € €possibilities. € solution. There may in fact be infinite If there are infinitely many possible linear combinations, then we can still give one particular such solution. € 4 Example 3. In R , u1 = (1, –2, 4, 2), u2 = (2, –4, 6, 3), and u3 = (5, –10, 16, 8). Determine if v = (–2, 4, –2, –1) is a linear combination of u1, u2 , and u3. € € € € Solution. Consider the associated system of equations: € € € 1 1 2 5 −2 −2 −4 −10 4 0 → rref → 6 16 −2 4 0 2 0 3 8 −1 0 1 0 0 1 2 0 0 4 −3 0 0 We see that there are infinitely many solutions: v = c1 u1 + c 2 u2 + c 3 u3 , where c1 = 4 − t ; c2 = −3 − 2t ; c3 = t , and t can be any real number. € 4, –2, –1) can be written as a linear combination of u , u , and u So the vector v = (–2, 1 2 3 in infinitely many ways. One particular solution (when t = 1) is c1 = 3, c 2 = –5, c3 = 1. Thus, v = 3 u1 – 5 u2 + u3. However the simplest solution is when t = 0. Then we have c1 = 4 and c2 = –3. € € € € Thus, v = 4 u1 – 3 u2 . € Note that the vectors u1 and u2 reduce to the 2 × 2 identity after rref. Thus, any € € that € can € be written as linear combination of u , u , and u can actually be written vector 1 2 3 as a linear combination of just and . u u 1 2 € € € € € € € € € € € Dr. Neal, WKU Exercises 3 1. In R , let u1 = (1, –2, –1) u2 = (–1, 3, 2) u3 = (2, –4, –2) u4 = (0, 2, 2) Let v = (11, 24, 10) and w = (11, –27, –16). € w are linear€combinations of€u1, u2 , u3, and u4 . If so, give the (a) € Determine if v and general form of all possible linear combinations as well as a specific linear combination. € € (b) Which are the only vectors really necessary to write any other vector as a linear € of €u , u , , and u ? € € € € combination 1 2 u3 4 2. In R 4 , let u1 = (1, –2, –1, 2) € € € € u5 = (1, –1, 0, 1) € € u2 = (–1, 3, 2, 3) u3 = (2, –4, –2, 4) u4 = (0, 2, 2, 0) and Let v = (1, 2, 3, 4) and w = (4, 3, 2, 1). € € € € (a) Determine if v and w are linear combinations of u1, u2 , u3, u4 , and u5. If so, give the general form of all possible linear combinations as well as a specific linear € € combination. € are €the only vectors really necessary € € to €write € any other € (b) Which vector as a linear combination of u1, u2 , u3, u4 , and u5? € € € € €