Formal Description of Dynamic Programming (DP)
Formal Description
si - denotes the state of the system at ith time, for all i = 1, 2, …, n (state transitions).
xi - denotes the decision made at ith time for all i = 1, 2, …, n.
Figure 1
In other words, for all i = 1,2,…,n-1, the state transitions are denoted as:
Let the cost of the changing state from si to si=1 be denoted as gi(si,xi)
Then the total cost is given as:
Design Methodology for DP
Four step method:
- Characterize the Bellman equation for the problem.
- Recursively find the optimal solution and itscorresponding value for each tail problem.
- Compute the value of an optimal solution in a bottom-upfashion.
- Construct an optimal solution from computedinformation.
The Squid