Linear dynamical system

A linear dynamical system is a mathematical model of how things change over time. In general, a dynamical system is a set of rules that describes how a state evolves. Most real-world systems are nonlinear, which makes them very difficult to solve with exact math. However, linear dynamical systems follow specific rules that make them much easier to study. They possess a rich set of mathematical properties that allow for exact solutions. Because of this, they are essential tools for understanding much more complex systems.

In these systems, we track changes using a state vector, which is an n-dimensional vector denoted as x. The change in this vector is determined by a constant matrix, denoted as A. This matrix acts on the current state to produce a variation. This variation can occur in two distinct ways. First, it can be a flow, where the state varies continuously with time. Second, it can be a mapping, where the state changes in discrete, separate steps. The term "linear" means that if you have two valid solutions, any linear combination of them is also a solution. This means you can combine solutions using scalars to find new ones.

Solving these systems often involves finding special directions within the matrix. If an initial vector aligns with a right eigenvector of matrix A, the movement is very simple. In this case, the dynamics are governed by a corresponding eigenvalue, denoted as lambda. The solution for such a vector is expressed as e^(lambda*t). If the matrix is diagonalizable, we can represent any vector in the n-dimensional space. We do this by using a linear combination of the right and left eigenvectors. This allows us to build a general solution for the entire system by combining individual solutions.

We can also use linear systems to study nonlinear ones through a process called linear approximation. Most nonlinear systems do not have closed-form solutions that are easy to calculate. However, we can find the equilibrium points, also known as fixed points, of a nonlinear system. Near these fixed points, we can approximate the messy nonlinear behavior with an equivalent linear system. This makes linear math a crucial first step for understanding complex, nonlinear worlds. Sometimes, a clever change of variables can even turn a nonlinear system into a linear one.

In a two-dimensional system, we can classify the behavior of fixed points using specific math tools. We use a characteristic polynomial to find the eigenvalues of the matrix A. The roots of this polynomial, which are the eigenvalues, determine the stability of the system. For a 2D system, the characteristic polynomial takes a specific form involving two values. These values are the trace, denoted as tau, and the determinant, denoted as delta. The trace is a single value from the matrix, while the determinant is another key number.

By looking at the relationship between these eigenvalues, we can predict the system's movement. The sign of the eigenvalues tells us if a fixed point is stable or unstable. If both eigenvalues are negative, the fixed point is considered stable. If both are positive, the point is unstable. If the eigenvalues have opposite signs, the fixed point is called a saddle. We can also use a discriminant to see if the point is nodal or a spiral. A spiral occurs when the eigenvalues are complex numbers.

These mathematical tools allow scientists to map out how systems will behave without seeing them move. By calculating the trace and the determinant, we can understand the qualitative behavior of a system. We can determine if a system will settle into a steady state or move away from it. This ability to classify points as stable, unstable, or saddles is vital for higher mathematics. Understanding these linear building blocks is the foundation for studying almost all dynamical processes in science.