! Problem Solving Processes
The following are some of the
ways of systematically approaching PROBLEMS or TASKS. Many
thinkers and authors have contributed to understanding this process that we all
do many times a day. We begin to learn our individual approaches very early in
life. Many of the ideas described below are common sense. It is useful
when communicating with others to use common descriptions of the processes.
Much of the following is often classified as: Systems
Analysis
Generally we are using the Scientific
Method which is often described in four steps:
1. Observation
- Observe whay happens; collect and study facts relevant to 'the problem'.
2. Hypothesis -
Explanation of the facts observed (and documented by data) by a mechanism,
theory, pattern, etc.
3. Prediction -
From two above make deductions, and predict what may happen given similar but
different circumstances.
4. Verification
- Test the results of two and three buy observing new data and comparing it
with predictions based on two and three.
Links
to other parts of this note:
- To section on Hall Matrix.
- To definition of TASKS
- To general view of a System Model
- To section on Degrees of Coupling"
The discussion below is intended to help
clarify thinking, understanding, and aid communication.
Problem solving usually involves:
1. Defining and describing the problem
2. Solutions Concepts - Choosing a solution strategy
3. Obtaining
domain and problem specific data
4. Modeling and Analysis of the effects of the solutions
5. Reporting and recommending a preferred solution
6. Implementing the preferred solution
A computer can be a useful
tool for many parts of the process, E.g.
for analysis, report preparation, data storage and retrieval. A computer can
replace some or all of:
- the paper, pencil, calculator operations and if used properly can
process data.
- (Manipulate, store, and recall information).
Using the computer
effectively (as for all work) requires good organization of the operations. One
effective description of the organization of a systems analysis is the Hall’s
MATRIX.
|
STRUCT. Fine ==> |
Problem |
Value |
System |
System |
Rank |
Decision |
Plan for |
|
Program Planning |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
Project Planning |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
System Development |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
Production |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
Distribution |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
Operations |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
|
Retirement |
-- |
-- |
-- |
-- |
-- |
-- |
-- |
Most projects and activities regardless of
size can be considered to have a set of activities that correspond to the
following sub-activities. Work tends to proceed from top left across the first row,
then to next row and so on until the lower right corner is reached. Each cell
may give rise to one or more tasks that may require organization similar to the
overall scheme. A project's Life Cycle theoretically spans from cell 1,1
to 7,7.
Life Cycle Analysis implies estimating all the cells of the matrix.
Ideally the analysis will be before undertaking a project. Such an analysis
assumes that it is possible to make estimates from conception to final disposal
that are sufficiently accurate to be meaningful. Great care is required to be
sure that the Futures used as the basis for estimating are neutral with
a high level of certainty. Neither overly pessimistic, nor optimistic estimates
are really useful. Audits or retrospective analyses are also often undertaken
to discover how the activities conformed to earlier plans and expectations.
Seven stages in the life of a project,
system, etc. are shown as rows. These are referred to as the COARSE
structure. Normally a decision maker, analyst, designer, etc. is working in one
of these stages. Within a stage seven activities or FINE stages are
identified. The ordering of the cells from left to right, top to bottom is a
logical in terms of when an activity should happen.
A project or item may have a very long life
so the time between cell 1,1 and 7,7 may be long enough to span more than one
human lifetime. On the other hand it may be very short and some of the implied
steps may be skipped.
Additional dimensions can be added to the
Hall Matrix. Some possible ones are listed below. This list has only been
partially tested and is not considered final. The items shown suggest the range
of desirable knowledge and available tools for systematic decision making,
analysis and design
Dimensions for the
Hall Matrix and possible topics:
1. Coarse Structure (as shown above).
2. Fine Structure
(as shown above).
3. TOOLS:
Data Analysis, Modeling, Operating, Reporting, Supervising, Utilities.
4. APPLICATIONS: Place, System, Institution, People, Economy, Nature (Environment),
Equipment
5. PROPERTIES:
Status, Structure, Inputs, Process, Outputs, Cycles, Linkages
The
HALL matrix or other coordinating structure can be used to:
a. Organize a
problem solving process
b. Identify
required techniques and information
c. Provide a storage and retrieval framework
d. Assist in scheduling the work
e. Information
for clients and workers about the phases and components of the work.
Another set
of the 'Basic Elements of Systematic Analysis':
Ref: Systems Analysis for Engineers & Managers; DeNeufville & Stafford;
p.6
TASKS
1. Definition of Objectives
2. Formulation of Measures of Effectiveness
3. Generation of Alternatives
4. Evaluation of Alternatives
5. Selection of Preferred Alternative
Evaluation involves the prediction of
expected performance.
A systems Analysis might be described as follows:
Determine the
STEPS:
1. Determine the
2. Specify and Calibrate a MODEL of the PROCESS.
3. Operate the MODEL with
4. Obtain ESTIMATES of the
Note: A MODEL may or may
not be reversible.
Inputs and Outputs may be random variables
A more general view of the
role of a system model in systematic analysis is shown below. The diagram
indicates that there are three sets of inputs: 'what is already happening',
'what can be controlled in the future' , and 'what can not'.
The multiple lines leading to and from the
model are intended to signify that there may be more than one flow path. The
products of the model are forecasts. These may be a range or distribution of
values rather than a single one.
Battelle Regional Forecasting Model PHASES:
Phase1. Socio-Economic Future forecast:
Population, Education, Income, Government spending patterns, Demand.
Phase2. INPUT-OUTPUT analysis of inter
industry relationships including the use of machinery to meet expected demand.
Technological change accounted for by changing tech coefficients.
Phase3. Forecasts of: WAGES, PROFITS,
PRICES, with introduction of changes in relative prices.
Some comments on notation:
It is usual for persons whose native language
is written from left to right is to use a 'left' >> 'right' convention
for their thinking and notation.
Decisions & Information (Data) about the
'problem' including the technology of 'HOW TO SOLVE IT' are supplied by the
'USER'. Good software will facilitate the supply of these inputs by the USER.
Smart software changes the kind of skill and knowledge required.
The software component can consist of any
type of information and set of instructions that are in machine readable code
and resident on machine readable media. The long term trend for software is to
provide the USER with more stored knowledge as in Expert Systems.
For a particular TASK there will usually be
some software defenses that must be overcome by the USER. The current state of
the art allows considerable variation in how to bridge the gaps.
Ideally the USER should be able to use the
hardware and software system as a consultant to supplement personal skill and
knowledge.
The USER should also be able to train a
software system so that it remembers the knowledge and data that are acquired
during processes. This intelligence should be available for future use by the
user and others with permission others linked in the user's network.
Most systems fall far short of these ideals.
Convenient 'Records' of USER supplied
information (data, knowledge, decisions) may be a peripheral activity to
problem solving but they should be incorporated in any process to the degree
that improves overall personal & institutional efficiency.
The minimal set of records is files of the
results. The next set of records is of the input data, and procedure. Beyond
these are records of operations, intermediate results, sources of information,
assumptions, weaknesses, and future use suggestions.
Complete records may become voluminous so
that careful organization of how they will grow and the recording media become
important.
Software and hardware continues to evolve.
One need to be able to effectively use the 'available' but should not loose
sight of future possibilities, especially if good records and an accumulating
knowledge base are preserved for future use.
No set of records is very useful without a
matched information search and retrieval system. The search & retrieval
should also grow with the records, preferably automatically as the records are
posted.
^ [NUNET] * [ Problem
Solving Process ] * [ Decision
Process, Decision Trees ] * [ Approaches to
Decision Making ]
End
to date: 061231 ams