Intelligent Systems And Their Societies Walter Fritz

Problem Solving

 

Problem solving is a technique developed by Kepner and Tregoe (Exterior link). They also call it "Analytic Trouble shooting". They explained the method in a very good book titled "Problem Solving and Decision Analysis". (This book seems hard to get. A similar book seems to be "The Rational Manager, Problem Solving and Decision Making ", By C.H.Kepner, available from amazon.com).

Problem solving is a technique that is particularly useful for solving problems at work, but it can also be useful for problems at home. Here we give a short summary of the main ideas.

 

Sudden Problems

Suppose that everything is running smoothly, but suddenly, with some equipment, some machine or tool, a problem occurs.
We call a problem an unwanted deviation from standard functioning. What we normally notice is an effect, something that happens and can be observed. This effect had a cause, that normally is invisible.
It is this cause that at some point in time produced a change, and so caused the effect, the deviation from standard functioning, that preoccupies us.

To correct the functioning we need to find the most probable cause.
First we have to define as well as possible what is the standard functioning and what is the present functioning. For this Kepner and Tregoe in their book "Problem Solving and Decision Analysis" recommend that we fill in a spreadsheet somewhat like this:

State the problem:
Data for checking a possible cause Data for stimulating thought
about a possible cause
What is wrong
in the way the object looks, sounds, feels, smells
What is NOT wrong
this time, even though
it is related, or was
wrong, at other times
Peculiarities
and differences
about the "wrong"
and "NOT wrong" columns
Changes
What, and at what
time, related with the "wrong" column
What is the object,
person, process
where you observe the defect or problem.
-
-
-
-
(in what object do you NOT observe the defect)
-
-
-
-
-
-
-
-
-
-
-
Where in the object
are the defects.
Where do you observe
the defective objects
-
-
-
-
(where do you NOT observe the defect)
-
-
-
-
-
-
-
-
-
-
-
When did the objects
or the defects at the
objects first appear.
Hour, circumstance?
-
-
-
-
(when did you NOT observe them)
-
-
-
-
-
-
-
-
-
-
-
How many or what
percentage of
defective objects, or
defects per object
do you observe.
Increasing? Size?
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

In the first column we fill in, with a few words, what precisely we can observe. A good definition of the problem is of great help in solving it.

The second column is of great importance. Here we look for contrast. If there are several machines and one has the problem, here we write down what is not wrong with the other machines, even though they could have the same problem. Or if it is just one machine, we try to contrast its problem with its prior functioning: what do we observe now but did not observe before. Or we could ask ourselves what was wrong sometime in the past, but is not the case now. Or what could reasonably go wrong, but is not the case now. Also we establish limits. The problem could be bigger or smaller, we state what we could have observed, but did not. The information in this column permits us to look for differences of all kinds.

In the third column we fill in peculiarities and differences, when looking at the first two columns. We write down what distinguishes that what we have written in the first column from that written in the second column (But we do not mention causes or effects).

In the fourth column we write down the change. When did each of the items, written down in the first column, change related to the peculiarity noted in the third column. Under what circumstances did it change.

When we have filled out the form completely we state all possible changes that could correspond with the data on this spreadsheet. These changes have to comply with the "wrong" column, the "NOT wrong" column, the "peculiarities" column and the "change" column. If a "change" does not correspond to these items, it is not one of the possible causes.
Having done that, we should be able to state the most probable cause. The cause is either the change itself or the change that permitted the cause to act.
The next step is to think: With this most probable cause, shouldn't we see other changes caused by it?
Now we look for these other changes. If we do observe them, then our most probable cause is right. We now should verify this, we should observe the probable cause.

 

Sudden Trouble with Persons

A problem with persons is a change, a deviation of the acting, of the performance, of a person, or a group of persons, from what is expected; and we would like to know the cause.
We use the same method described under "Sudden Problems"; but it is more difficult to work with people than with inanimate objects. The following comments will help you in analyzing problems with persons. Problems with people often are not a single problem but a list of problems. Analyze the biggest problem first.

It is nearly impossible to suddenly change the character of a person, so we work by changing the environment.

Here are some typical difficulties when working with persons, which we should avoid:

When stating the "problem", there is a tendency to state conclusions or opinions and not the acts that have been observed. For instance: John has lost interest in his work (instead, we should state specifically and in detail, what is observable, what he does wrong).
The incorrect statement of the problem or an incorrect entry in the "wrong" column will make it very hard to find the change and then the cause.
In the "peculiarities" column we should only state differences related to the environment, not differences of the person whose activity we are analyzing. Other persons may be part of the environment of the first person. If they acted different, that is a correct "peculiarity".
Often persons do not react immediately to changes, sometimes they think about the changes and react later; sometimes only an accumulation of changes makes them act different. With people we should look for changes that occurred some time before the effect (the acting) was noted.
Persons are not things. Therefore we should be careful before we try corrections. We should be reasonably certain that we have the real cause and a good way of correcting it (without causing new problems).
Resolving problems with persons is not easy, but if we analyze unemotionally, and based on observable facts, often we will be successful.

All this is a very short summary. To apply this technique well, a training course is needed. See the training courses at Kepner Tregoe Worldwide (Exterior link).

 

For continuous reading, like a book - continue here.
Jump to the How to increase our Intelligence document / e-book Contents.


Last Edited 3 June 2013 / Walter Fritz
Copyright © New Horizons Press