To find maxima or minima of functions of two variables, one must find the critical points by setting the first partial derivatives equal to zero, i.e., fx=0 and fy=0. Then, the second derivative test is applied by computing D=fxx fyy - (fxy)². If D>0 and fxx>0, there's a local minimum; if D>0 and fxx<0, there's a local maximum. A saddle point occurs if D<0, and if D=0, the test is inconclusive. A saddle point is a point where the function changes curvature, resembling a saddle shape, which is neither a max nor min. For example, for f(x,y)=x²+y², (0,0) is a minimum because D=fxx fyy-(fxy)²=4-0=4>0 and fxx=2>0 .

Taylor series expansion for functions of two variables allows approximation around a point (a, b). It is given by f(x,y) ≈ f(a,b) + (x-a)fx(a,b) + (y-b)fy(a,b) + ½[(x-a)²fxx(a,b) + 2(x-a)(y-b)fxy(a,b) + (y-b)²fyy(a,b)] + ... This expansion utilizes the function's value, its first derivatives, and second-order partial derivatives at the point (a, b), enabling an estimate using a polynomial. This method outputs a quadratic approximation representing the nearby behavior of the function, significantly useful when analyzing local properties .

Maclaurin series expansions are specific cases of Taylor series where the expansion is around zero. These expansions efficiently approximate elementary functions using power series. For instance, e^x is expanded as 1 + x + x²/2! + x³/3! + ...; sin x as x - x³/3! + x⁵/5! - ...; and cos x as 1 - x²/2! + x⁴/4! - .... The patterns inherent in these power series derive from repeatedly differentiating the function at zero and are exemplary forms showcasing the convergence of Taylor series into straightforward expressions for key functions in calculus .

Consider the function z=x²y where both x and y are time-dependent variables, such that x=t and y=t². To find the total derivative with respect to time, first compute the partial derivatives: ∂z/∂x = 2xy and ∂z/∂y = x². Then, substitute the expressions for x and y, and calculate the derivatives of x and y with respect to t: dx/dt=1 and dy/dt=2t. Thus, dz/dt = (2xy)(1) + (x²)(2t) = 2t·t² + t²·2t = 4t³ .

The Jacobian determinant plays a crucial role in changing variables within multiple integrals, effectively managing the transformation of coordinate systems. During such transformations, the Jacobian accounts for how the volume element scales under the variable change. For example, when converting Cartesian coordinates to polar coordinates, the Jacobian incorporates how area changes are reflected when switching systems, computed as the determinant of the matrix of partial derivatives of the new variables with respect to the old ones. This allows accurate calculation of integrals in the new coordinate system by correcting for the area distortion introduced by the transformation .

Partial differentiation involves computing the derivative of a function with respect to one variable while keeping other variables constant, thus analyzing the effect of that specific variable. For example, for z=f(x,y), the partial derivatives would be ∂z/∂x and ∂z/∂y, each focusing on one variable. Total differentiation, however, considers how a function's output changes with respect to all variables changing simultaneously, often using time as a parameter. In this case, if z=f(x,y), x and y being functions of t, then dz/dt can be calculated using the chain rule: dz/dt = (∂z/∂x)(dx/dt) + (∂z/∂y)(dy/dt).

Both Taylor and Maclaurin series are used to approximate functions using polynomials. The Taylor series expands a function around any point a, given by f(x) = f(a) + (x-a)f'(a) + (x-a)²/2! f''(a) + ..., whereas the Maclaurin series is a special case of Taylor series expanded around a=0. Therefore, the series becomes f(x) = f(0) + xf'(0) + x²/2! f''(0) + ... These expansions allow complex functions to be represented as infinite sums of polynomials, facilitating easier computation and analysis. For instance, the Maclaurin series for e^x is 1+x+x²/2!+... The primary distinction is the point around which the function is expanded .

The chain rule in the context of several variables is used to differentiate composite functions. If a function z depends on a variable u, which in turn is a function of x and y (u=g(x,y)), the chain rule allows us to express the derivative of z with respect to x as ∂z/∂x = (dz/du)(∂u/∂x). An example is z=sin(u), where u=x²+y, leading to ∂z/∂x = cos(u)·(2x) = 2x cos(x²+y).

Taylor's series allows representation of differential equations through series expansion, simplifying complex equations into polynomial approximations. These polynomial forms facilitate numerical methods like Euler's method or Runge-Kutta, which solve differential equations by iterative processes. In essence, Taylor expansions provide the incremental approach needed for efficient numerical integration by offering initial slope values and approximations that significantly enhance accuracy and convergence of algorithms evaluating differential equations .

The Jacobian is a determinant used to represent the best linear approximation of a differentiable function near a point. It is computed for functions that map R^n to R^m and involves partial derivatives of all functions involved. Specifically, for functions u=f(x,y) and v=g(x,y), the Jacobian J is given by the determinant | ∂u/∂x ∂u/∂y ; ∂v/∂x ∂v/∂y |. The Jacobian not only helps in transformation of coordinates within integrals, particularly in relation to multiple integrations, but also plays a key role in analyzing the local behavior of the function, such as checking if a function is invertible at a point .