Linearity

A system is linear if it can be defined by linear differential equations. In particulars, if the functions and in its state-space model are linear functions of the state variables and input .

Linear systems have two properties:

  1. Homogeneity: Scalar multiplication works as we expect.
    • If we have a system that takes as input and outputs , inputting will instead return .
  2. Additivity/superposition: Addition works as we expect it to.
    • If we have a system that has and , passing in will result in

Time-Invariance

A system is said time-invariant if it can expressed be by differential equations with constant coefficients. In particular, if the functions and in its state space model do not depend on the time explicitly.

Assume that a time-invariant system has zero initial conditions, and zero input generates zero output. If input produces output , then input for all .

So if a system has , but instead we input , we will have

If , then the system is time invariant.

Examples

Example 1

Let’s say we have . Then, we have

Since the two are not equal, the system is not time-invariant.

However, we have

so this system is linear.

Example 2

Let’s say we have . Then:

Thus, this is time invariant.

However, we have

Thus this system is not linear.

Example 3

Let’s say we have

Because of the term, this is time-varying.

Example 4

Let’s say we have

Because of the term, we know that this is not linear (similar to Example 1).

Form

A LTI system has the following form of state space model:

where , , , and are constant matrices.

State Machine Interpretation

Linear time-invariant systems can be seen as a state machine where , and and are linear functions of their input. In discrete time, they can be described as a linear difference equation, like

where is at time . LTI systems can be implemented using state to store relevant previous input and output information.

Recurrent Neural Networks are a lot like a non-linear version of LTIs.