Dynamic Programming (DP)

~Concept, Properties and Principle of Optimality

~Stagecoach Problem

Prepared by: Kevin DO. Roa

Concept/Introduction
Consider a company that has to decide on the production plan of an
item for the next three months so as to meet the demands in different
months at minimum cost. The different months for which the production is
to be decided constitute the stages. So it is a multi-stage problem. For such
a problem, decisions are made sequentially over several periods using a
technique called Dynamic Programming (DP) technique. One thing
common to all models in this category is that current decisions influence
both present and future periods.

The dynamic programming approach divides the problem into several

sub-problems or stages and then these sub-problems are solved
sequentially until the initial problem is finally solved. The common
characteristic of all dynamic programming models is expressing the
decision problem by means of recursive formulation.
Dynamic Programming
Dynamic programming is a useful mathematical
technique for making a sequence of interrelated
decisions. It provides a systematic procedure for
determining the optimal combination of decisions.

Incontrast to linear programming, there does not

exist a standard mathematical formulation of “the”
dynamic programming problem. Rather, dynamic
programming is a general type of approach to
problem solving, and the particular equations used
must be developed to fit each situation.
Terminology
Terminologies which are commonly used in dynamic
programming are given below:

Stage - The point at which a decision is made is known as a stage.

The end of a stage marks the beginning of the immediate
succeeding stage. (For instance, in the salesmen allocation
problem, each territory represents a stage; in the shortest route
problem, each city represents a stage.)

State - The variable that links two stages in a multistage decision

problem is called a state variable. At any stage, the values that
state variables can take describe the status of the problem. These
values are referred to as states. (For example, in the shortest route
problem, a city is referred to as state variable.)
Principle of optimality: The principle of
optimality states that the optimal decision
from any state in a stage to the end, is
independent of how one actually arrives at
that state.

Optimal policy: A policy which optimizes the

value of an objective function is called an
optimal policy.
Bellman’s principle of optimality : It states that
“an optimal policy (a sequence of decisions) has the
property that whatever the initial state and decisions
are, the remaining decisions must constitute an
optimal policy with regard to the state resulting
from the first decision.”

Return function : At each stage, a decision is made

which can affect the state of the system at the next
stage and help in arriving at the optimal solution at
the current stage. Every decision has its merit which
can be represented in an algebraic equation form.
This equation is called a return function.
Characteristics of DP
The basic features which characterize the dynamic
programming problem are as follows:

(a) The problem can be subdivided into stages with a

policy decision required at each stage. A stage is a
device to sequence the decisions. That is, it
decomposes a problem into sub-problems such that an
optimal solution to the problem can be obtained from
the optimal solutions to the sub-problems.

(b) Every stage consists of a number of states

associated with it.
(c) Decision at each stage converts the current stage into state
associated with the next stage.

(d) The state of the system at a stage is described by a set of

variables, called state variables.

(e) When the current state is known, an optimal policy for the
remaining stages is independent of the policy of the previous ones.

(f) To identify the optimal policy for each state of the system, a
recursive equation is formulated with n stages remaining, given the
optimal policy for each state with (n − 1) stages left.

(g) Using recursive equation approach each time, the solution

procedure moves backward stage by stage for obtaining the
optimum policy of each state for that particular stage, till it attains
the optimum policy beginning at the initial stage.
The Stagecoach Problem

Image source: https://www.stahlsauto.com/automobiles/1860-

stagecoach/
The Stagecoach Problem
The STAGECOACH PROBLEM is a problem specially
constructed to illustrate the features and to introduce the
terminology of dynamic programming.

“It concerns a mythical fortune seeker in Missouri who

decided to go west to join the gold rush in California
during the mid-19th century. The journey would require
traveling by stagecoach through unsettled country where
there was serious danger of attack by marauders.
Although his starting point and destination were fixed,
he had considerable choice as to which states (or
territories that subsequently became states) to travel
through en route.”
The possible routes are shown in Fig. 11.1, where
each state is represented by a circled letter and the
direction of travel is always from left to right in the
diagram. Thus, four stages (stagecoach runs) were
required to travel from his point of embarkation in
state A (Missouri) to his destination in state J
(California).
This fortune seeker was a prudent man who
was quite concerned about his safety. After
some thought, he came up with a rather clever
way of determining the safest route.

Life insurance policies were offered to

stagecoach passengers. Because the cost of the
policy for taking any given stagecoach run
was based on a careful evaluation of the safety
of that run, the safest route should be the one
with the cheapest total life insurance policy.
The cost for the standard policy on the
stagecoach run from state i to state j, which
will be denoted by cij , is

These costs are also shown in Fig 11.1.

We shall now focus on the question of
which route minimizes the total cost of the
policy.
Solving the Problem
Solving the Problem
First note that the shortsighted approach of selecting the
cheapest run offered by each successive stage need not
yield an overall optimal decision.

Following this strategy would give the route A->B->F->

I->J, at a total cost of 13. However, sacrificing a little on
one stage may permit greater savings thereafter. For
example, A->D->F is cheaper overall than A->B->F.

One possible approach to solving this problem is to use

trial and error. However, the number of possible routes is
large (18), and having to calculate the total cost for each
route is not an appealing task.
Fortunately, dynamic programming provides a
solution with much less effort than exhaustive
enumeration.

Dynamic programming starts with a small

portion of the original problem and finds the
optimal solution for this smaller problem. It
then gradually enlarges the problem, finding
the current optimal solution from the
preceding one, until the original problem is
solved in its entirety.
For the stagecoach problem, we start with the smaller problem
where the fortune seeker has nearly completed his journey and
has only one more stage (stagecoach run) to go.

The obvious optimal solution for this smaller problem is to go

from his current state (whatever it is) to his ultimate destination
(state J).

At each subsequent iteration, the problem is enlarged by

increasing by 1 the number of stages left to go to complete the
journey.

For this enlarged problem, the optimal solution for where to go

next from each possible state can be found relatively easily
from the results obtained at the preceding iteration. The details
involved in implementing this approach follow.
Solution Procedure
Characteristics of DP Problems
The stagecoach problem is a literal prototype of
dynamic programming problems. In fact, this
example was purposely designed to provide a
literal physical interpretation of the rather
abstract structure of such problems.

Therefore, one way to recognize a situation that

can be formulated as a dynamic programming
problem is to notice that its basic structure is
analogous to the stagecoach problem.
These basic features that characterize dynamic
programming problems are presented and discussed here.

1. The problem can be divided into stages, with a policy

decision required at each stage.
2. Each stage has a number of states associated with the
beginning of that stage.
3. The effect of the policy decision at each stage is to
transform the current state to a state associated with the
beginning of the next stage (possibly according to a
probability distribution).
4. The solution procedure is designed to find an optimal
policy for the overall problem, (i.e., a prescription of the
optimal policy decision at each stage for each of the
possible states.)
5. Given the current state, an optimal policy
for the remaining stages is independent of the
policy decisions adopted in previous stages.

6. The solution procedure begins by finding

the optimal policy for the last stage.
 References/Sources
Giri, B. C. (n.d.) Operations Research.
Chapter 8: Dynamic Programming.
Department of Mathematics, Jadavpur
University. Kolkata, India. Pdf.

Chapter 11: Dynamic Programming (n.d.)

Retrieved from https://www.ime.unicamp.br/~
andreani/MS515/capitulo7.pdf