Problem Understanding and Analysis

What is a problem?
• There are many different kinds of problems:

• How much tax should my employees be paying?

• How do I get to Oxford from London?

• Why is it wrong to lie?

Cont..
• However, not all kinds of problems can be solved using computers
(Which of the above do you think can and cannot be solved with a
computer?).

• For a problem to be solvable through the use of a computer, it must

generally:

• Be technical in nature with objective answers that can be

arrived at using rational methods
• Be well-structured
• Contain sufficient information for a solution to be found
• Contain little, if any, ambiguity.
How do we solve problems?
• We need to THINK!

• We need to engage in one or more of the following types of thinking:

• Logical reasoning (e.g. Dave and Anne have two daughters. Each
daughter has a brother. How many people are there in the family?).

• Mathematical reasoning (e.g. A car has a fuel capacity of 10 gallons

and a fuel consumption of 45 miles per gallon. How far can the car travel on
three quarters of a tank?).

• Lateral thinking (e.g. I throw a heavy weight to arrive at a district in

North London?).
How do we solve problems?

• Problem solving is easier if we employ a

problem solving framework and an
appropriate problem solving strategy to
aid us.
A problem solving framework
• An outline framework for problem solving:
1.Understand the problem
2.Devise a plan to solve the problem
3.Carry out the plan
4.Assess the result
5.Describe what has been learned from the process
6.Document the solution.
• This framework is called How to Think Like a
Programmer (HTTLAP) approach to problem solving.
(1) Understanding the problem
• To understand a problem

• We need to read and reread it till we understand every

detail
• We need to dissect the problem into its component parts
(e.g. problems and sub-problems)
• We need to remove any ambiguity
• We need to remove any information that is extraneous to
the problem
• We need to determine our knowns and our unknowns
• We need to be aware of any assumptions we are making.
(2) Devise a plan to the solve the problem
• If a problem contains a set of sub-problems, in what order are you going to solve
them?

• How are you going to represent the problem:

• Numerically?
• Graphically?
• Tabular data?
• Natural language?

• Does the problem lend itself to a particular problem solving strategy or

strategies:
• Working backwards?
• Solving a simpler analogous problem?
• Logical reasoning?
• Finding a pattern?
(3) Carry out the plan
• Consider the following problem:
In a room with ten people, everyone shakes hands with
everyone else exactly once. In total, how many
handshakes are there?
• How would you represent and solve the problem?
• To solve it quickly and correctly without guessing we
could use one of the strategies mentioned earlier.
However, which strategy is best suited to solving this
problem?
• To answer this, we need to familiarize ourselves with the
strategies and practice their use.
Cont..
• To answer the handshakes problem we could begin by
accounting for all the possibilities and graphically A B C D E F G H I J
represent those possibilities in a table. A X
B X
• This clearly shows us who shakes hand with who (A to C X
J shake hands with everybody except themselves (X)). D X

• We can then use some basic mathematical reasoning E X

to find our solution: F X

G X
H X
number of people = 10 I X

grid size = 10 x 10 = 100 J X

handshakes = (100 - 10)/2 = 45

• From here we can abstract a generalisable solution

that can easily be turned into an algorithm and
eventually into computer code:
h = ((p2) – p))/2
(4) Assessing the results
• It is very unusual when solving complex problems to achieve the
correct result first time round. We often need several attempts to
get it right.
• To verify our solutions are correct, we need to take a few steps
backwards:
• Was our understanding of the problem correct?
• Did we overlook anything?
• Did we choose the correct strategy?
• Did we employ that strategy correctly?
• Have we made any incorrect or unwitting assumptions?
• However, it is often very difficult to spot our own mistakes. It is
often better, therefore, to have somebody else verify our solutions
for us.
Cont...
• Sometimes solutions appear correct, but are in fact
wrong, due to an initial misunderstanding of a
problem
• If you have misunderstood a problem, it does not
matter how good a coder you are, your program
will not work as it is supposed to.
• Therefore getting the problem-solving part of
programming right is absolutely essential if we are
to build programs that work as they are supposed
to work.
(5) Describing what you have learned
• You can only become a good problem solver by reflecting
on your experiences of problem solving.

• Keeping a record of problems you have attempted, your

success, failures, the approaches you have used, etc. will:
• Broaden your problem solving repertoire
• Help you to recognize similarities/patterns in problems
• Help you identify and fix logical or implementation
errors
• Help you to solve problems faster and more effectively
(6) Documenting the solution
• Documenting a solution will help you to generalize your
approach to other similar problems.
• Do you recognize the following type of problem? How would
you solve it?

• It will also prevent you forgetting how you arrived at your

solutions.

• In the long run, it will make you a better problem solver

and eventually a better programmer.
Strategies Strategies for Problem Solving
for
arriving at Problem solving refers to the thinking we do in
solutions order to answer a complex question or to figure
include: out how to resolve an unfavorable situation.
Trial and error involves trying various possible solutions, and
if that fails, trying others.
trial and •When it’s useful: perfecting an invention like the light bulb
error by trying a thousand filaments
•When it fails: when there is a clear solution but trial and error
might miss it forever
algorithms An algorithm is a step by step strategy for solving a problem,
methodically leading to a specific solution.
A heuristic is a short-cut, step-saving thinking strategy
heuristics or principle which generates a solution quickly (but
possibly in error).

insight Insight refers to a sudden realization, a leap forward in

thinking, that leads to a solution.
Clarifying Problem Solving Examples
Where’s Tothefind
apple
a
juice? Dospecific
I look on every
item in
shelf in the store, or do I
a supermarket
search where there is
similar stuff?

Trial and Wander around a

error supermarket
randomly to find it.

Create a methodical
Algorithms path to make sure
you check every
single aisle.

Heuristics Check only

related aisles.
End of Lecture 9