Now, let me introduce to you another method of finding a particular solution for non-homogeneous linear system of equations. The so-called the variation of parameters. Variation of parameters. Let's go back to our original problem. Say the system of equations, X prime is equal to A times X plus the possible non-homogeneous term F. The general theory said, it's general solution, consists of the two parties, the complimentary solution plus the particular solution. We have introduced the method of undetermined coefficients for finding this particular solution, Xp, but in the method of undetermined coefficients, they allow over top of this non-homogeneous term, capital F is rather restricted way. Now, we will discuss the another method, the method of variation of parameters to find this Xp, a particular solution. What is Xc here? We know that the complimentary solution Xc. That is of the form of the summation of j is equal to one to the n, the arbitrary coefficient Cj times Xj. That's the complimentary solution, Where the collection Xj. Collection Xj is a so-called the fundamental set of solutions. This is the so-called the fundamental set of solutions for corresponding homogeneous problem, say X prime is equal to A times X. In other words those Xj's are, they are n linearly independent solutions of X prime is equal to A times X. That's the form of the complimentary solution X_c. Now, let's introduce the convenient the matrix form of this general solution of the corresponding homogeneous problem. Look at the following X_c, that is the sum of the linear combination of n linearly independent solutions of the corresponding homogeneous problem, Cj times Xj. We can write it in the following way. This is the same as matrix having n columns given by the X_1, X_2 and X_n because it is checks j is the column vector of dimension n. This is a n by n matrix. As you can see, this is n by m matrix, times the column vector C_1, C_2, dot and the C_n. This linear combination, we can write it in this fashion. Think about the ray. It's easy matter to figure out this matrix product is the same as that one. Then I will write it as a capital V of t Palms a capital C. Where what is the capital P of t? As you can see, this is the n by n matrix, having n columns from X_1, X_2 through the X_n. That is the capital V of t. What is a capital C? This is just the column vector C_1, C_2 through the C_n. This n by m matrix having those extra Xj's, as is the column vector. This is the call that we call it as the fundamental matrix. For homogeneous problem X' is equal to A times X. As you can see, each C_j is an arbitrary constant. Capital C is an arbitrary constant vector. This is an arbitrary constant vector. In other words, this complimentary is always an X_c. We can simply write it, X_c is equal to V_t times C. Fundamental matrix times the arbitrary, the constant vector. That's the form of the complimentary solution. Now, the essential part of the method of variation of parameters is the following. Replace this arbitrary constant vector C by some unknown column vector of functions. That's the starting point. Before discussing the top of the particular solution X_p, let's make one important comment. I claim that, note to the following, it's very important. Note that capital V of t, this is an n by n matrix so that we can consider its determinant. I'm claiming that this is never zero. In other words, when the n by n matrix has a non-zero determinant, we say that the n by n matrix is non-singular. In other words, we can say that note that the fundamental matrix V of t is a non-singular for any T everywhere. Can you see why? Remind the following. What are the columns of the fundamental matrix, its component columns are X_1 and X_2 and through the X_n and they are linearly independent solutions, linear independent. That's why this fundamental matrix is non-singular everywhere. Non-singular means what? It implies the n-by-n matrix always has its inverse. Not every square matrix has an inverse, the any n by n matrix has an inverse if and only if the matrix is non-singular. Because this is non-singular, that implies that it has an inverse. That's the key point. Now, it's time to make the initial guess for the particular solution. Now let me erase this part. We expecting. Make a guess. Our guess is that there is a particular solution X_p, which is the tie V of t. Remind what is X_c? X_c is a fundamental matrix times an arbitrary constant vector. Variational parameters. We'd like to change this arbitrary constant of the vector by a vector of unknown functions. Like I said X_p is equal to V times U of t. What is the U of t? U of t, this is a column vector having n components, U and t, U to t, and the U_n of t.This is unknown vector of functions. We're guessing that there must be a solution. There must be a particular solution of this non-homogeneous problem of type fundamental matrix times the sum, the column vector of unknown functions. Let's just say that actually the possibility, whether there is a really a particular solution of this form or not. That's easy. Plugging this one into that. If this is a solution to this problem then, X_p prime is equal to v prime of t times the u of t plus v of t times u of t, by the product rule. Where you must note the following. What's the derivative of V fundamental matrix? This is, it has a component, x_1 prime and x_2 prime, and x_n prime. What are those extra j's? Extra j's are solutions of corresponding homogeneous problem, so that this is equal to C_1. No, x_1 prime is equal 8 times x_1, because it's the solution to the corresponding homogeneous problem. X_2 prime is equal to 8 times x_2 and so on. X_n prime is equal to 8 times x_n prime. That is equal to. You can write this as A times V prime again. Do you agree to it? So that using this identity that is equal to 8 times, no, I made a mistake. Let me say it again. X_1 prime is equal to ax_1, x_2 prime is equal to ax_2, x_n prime is equal to a times x_n. This is equal to just the a times the V, so a_V times U plus V times the U. That must be equal to the right-hand side. What is the right-hand side? When x is equal to X_p, a times X_p is equal to a_V times V times U, and plus the F. That's the right-hand side. I think I made some minor mistake here. The X_p prime is apply the product rule, V prime u plus V times u prime of V. That's it. This is the right-hand side of the differential equation. We're almost ready to conclude. Compare the two sides of the equation. If there is a particular solution of this type then, we must have this equality. From this, these two coincide, so canceled out. Then means it's a solution. X_p is a solution to this original problem, if and only if V times u, that is equal to F. I'm sorry, V times u prime. I forgot to put the prime here. Vu prime. This is a solution to this problem. Then we must have that word. In fact, what I'm claiming is, let's make my claim the more precise. Through this computation we can say then, X_p, V times U, this is a solution of our original problem x prime is equal to a times x plus f, if and only if P times a U prime that is equal to f. Because I'm saying, this fundamental matrix is always non-singular. It's inverse. That means U prime is equal to V inverse times F.