Summary
Highlights
The video shows how to simplify the system of equations. It's observed that both equations are equivalent, meaning one equation is redundant. An additional crucial piece of information is used: since p and q represent probabilities, their sum must always be 1 (p + q = 1).
The simplified equation (e.g., Q = 2P) is combined with the condition P + Q = 1. By substituting one variable into the other equation, the values for P and Q are determined. In this example, Q is found to be 2/3 and P is 1/3.
The video concludes by stating that the stable state for this graph is P = (1/3, 2/3). This means that after a very large number of steps, the probability of being in state A is 1/3, and the probability of being in state B is 2/3, regardless of the initial state.
The video demonstrates how to multiply the row matrix P (p, q) by the transition matrix M. This product yields a new row matrix whose elements are expressions involving p and q.
The video introduces the concept of a stable state in an oriented, weighted probabilistic graph. After a very large number of steps, the probability of being in a certain state (e.g., A or B) tends towards a constant probability that no longer changes.
To determine the stable state, a transition matrix (M) is needed. This matrix translates the probabilities of moving between states into a numerical table. For a graph with two vertices A and B, the matrix M will represent the probabilities of transitioning from A to A, A to B, B to A, and B to B.
The stable state, denoted as 'P' (with coefficients 'p' and 'q'), is defined by the formula P = PM. This means that after a stable state is reached, multiplying the state vector P by the transition matrix M results in the same state vector P.
Since P = PM, the resulting matrix from the multiplication is set equal to the original state matrix P. This leads to a system of two equations with two unknowns (p and q) because two matrices are equal if and only if their corresponding coefficients are equal.