We can now look at the field of Systems Engineering and
see that concepts can be applied to the understanding of the proposed heuristic.
p.xi"This book is about systems. It focuses on the engineering
of systems and on systems analysis."
Fabrycky and Blanchard teach that systems have components, attributes, and relationships.
This is much like the situation on a chessboard where we have pieces, that have defined moves, and
which interact with the other pieces on the board. The parts of a system work together to achieve a common purpose.
Each part of a system contributes to the success of the whole. The pieces on a chessboard work together to achieve a
common purpose.
p.3"The Elements of a System [section title] Systems are composed of components, attributes, and
relationships. These are described as follows: 1. Components are the operating parts of a system consisting of input,
process, and output. Each system component may assume a variety of values to describe a system state as set by some control
action and one or more restrictions. 2. Attributes are the properties of discernible manifestations of the components
of a system. These attributes characterize the system. 3. Relationships are the links between components and attributes.
A system is a set of interrelated components working together toward some common objective or purpose."
A chess game does not have to take on a physical form, as in pieces and a board. We can define
all the necessary elements of the game using symbols. A chess game might be categorized as a conceptual system.
p.7"Physical systems are those that manifest themselves in physical form. They are composed
of real components and may be contrasted with conceptual systems, where symbols represent the attributes of components."
Engineering activities are directed towards satisfying human wants. There appears to be a need for a software
tool that generates accurate analysis of chess games, including positional factors which contribute to the chances of winning.
p.16"Engineering activities of analysis and design for human-made or technical systems are not an end in
themselves but are a means for satisfying human wants. Thus, modern engineering has two aspects. One aspect concerns itself
with the materials and forces of nature; the other is concerned with the needs of people."
What is the definition of Systems engineering? Here is one definition:
p.18"[Systems Engineering is] An approach to translate operational needs
and requirements into operationally suitable blocks of systems. The approach shall consist of a top-down, iterative process
of requirements analysis, functional analysis and allocation, design synthesis and verification, and system analysis and control.
Systems engineering shall permeate design, manufacturing, test and evaluation, and support of the product. Systems engineering
principles shall influence the balance between performance, risk, cost, and schedule."
New systems have their foundation in the needs, wants or desires which emerge from
the operation of a current system. If a desire is strong enough, organizations begin to plan building new systems
which better meet the needs of users. What if we applied the principles of systems engineering to the problem
of building a computer chess program to play a strong positional game of chess? We would start at the place we start
for all systems - with the identification of need. We might decide that the current set of computer chess programs
do not have adequate insight into positional factors which contribute towards winning chances. We might therefore look at
or propose alternative architectures which give us this insight.
p.55-58"PROBLEM DEFINITION AND NEED IDENTIFICATION [section title] The systems engineering
process generally commences with the identification of a 'want' or 'desire' for something and is based on a
real (or perceived) deficiency. For instance, suppose that a current system capability is not adequate in terms of meeting
certain required performance goals, is not available when needed, cannot be properly supported, or is too costly to operate...
Defining the problem is sometimes the most difficult part of the process... Having justified the need for a new system,
it is then necessary to (1) identify various possible system-level design approaches that could be pursued in response to
the need; (2) evaluate the most likely approaches in terms of performance, effectiveness, maintenance and sustaining
support, and life-cycle economic criteria; and (3) recommend a preferred course of action. There may be many possible alternatives;
however, the number must be narrowed down to a few feasible ones, consistent with the schedule requirements and the availability
of resources"
The model that we create to represent the real world for the proposed system is a critical
component of our computer program. The degree of our success or failure will depend on what we choose to represent, as well
as what we choose not to represent, in this model. Blanchard and Fabrycky feel that a model of a system should
include all of the following:
p.109"A model, in this context, is a simplified representation of the real world which abstracts
features of the situation relative to the problem being analyzed... The model used should incorporate the following features.
1. The model should represent the dynamics of the system configuration being evaluated in a way
that is simple enough to understand and manipulate, and yet close enough to the operating reality to yield successful results.
2. The model should highlight those factors that are most relevant to the problem at hand, and suppress
(with discretion) those that are not as important.
3.The model should be comprehensive by including all relevant factors and be reliable in terms of
repeatability of results.
4.Model design should be simple enough to allow for timely implementation in problem solving...
5.Model design should incorporate provisions for ease of modification or expansion to permit the
evaluation of additional factors as required."
Blanchard and Fabrycky speak of systems having objectives, and of the limiting factors
that exist which threaten (or actually prevent ) our system from achieving its objectives. Strategic points can be
creatively identified that, when altered, can make progress towards the objectives possible. A great concept for our computer
chess program. They also speak of the advantages of simplifying complex real-world situations by using a model. Another great
idea.
The following quotation is an outstanding explanation of the concepts involved in the proposed
heuristic. Objectives, limiting factors, and strategic factors are all used to focus the search
efforts and to identify and examine move sequences that are interesting and have the most potential to 1) be lines
of best play or 2) provide additional insight in selecting one move over others.
p.162-163"An
understanding of the decision-making process usually requires simplification of the complexity facing the decision maker.
Conceptual simplifications of reality, or models, are a means to this end... Those factors that stand in the way of attaining
objectives are known as limiting factors. An important element of the systems engineering process is the identification
of the limiting factors restricting accomplishment of a desired objective. Once the limiting factors have been identified,
they are examined to locate strategic factors, those factors that can be altered to make progress possible.
The identification
of strategic factors is important, for it allows the decision maker to concentrate effort on those areas in which success
is obtainable. This may require inventive ability, or the ability to put known things together in new combinations, and is
distinctly creative in character. The means that will achieve the desired objective may consist of a procedure, a technical
process, or a mechanical, organizational, or managerial change. Strategic factors limiting success may be circumvented by
operating on engineering, human, and economic factors individually and jointly."
To compare alternatives, we must convert the different kinds of perceived advantage (as well as
intangibles such as risk and redundancy) to a common factor. We cannot compare apples to oranges. Ideally we desire an estimate
of winning chances, and we select the move with the highest estimated winning chances.
p.164"To compare alternatives equivalently, it is important that they be converted to a common measure.
This conversion to a common measure permits comparison on the basis of equivalence.... decisions between alternatives should
be made on the basis of their differences. Thus, all identical factors can be canceled out for the comparison of any two or
more alternatives at any step of a decision-making process."
Creating a model of the situation lets us perform thought experiments and 'play' with the situation
as it presently exists. Blanchard and Fabrycky emphasize that essential relationships must be included
in the system model. In the proposed heuristic, this would be the relationships among the pieces on the board and
their potential ability to accomplish objectives and to potentially deny the opponent's pieces that same ability.
p.164-165"Models and their manipulation (the process of simulation) are useful tools in systems
analysis... Models are designed to represent a system under study, by an idealized example of reality, to explain the essential
relationships involved."
A model used to make a decision should include the interactions that are significant and
which contribute significantly to the outcome.
p.168"Because of the impossibility of including all factors in constructing the evaluation function
[Blanchard and Fabrycky are speaking of a general mathematical decision model], it is common practice to consider only those
on which the outcome is believed to depend significantly... Although it may be feasible to consider only those relationships
that are significantly pertinent, one should remember that all system elements are interdependent."
When making a decision, we need to form a common basis so we can rank and then choose among the
alternatives.
p.169"Decision evaluation is an important part of systems engineering and analysis. Evaluation is
needed as a basis for choice among alternatives that arise from design activities, as well as for optimizing systems already
in operation. In either case, equivalence provides the common evaluation measure on which choice can be based."
When trying to predict the future, it is human nature to consider the probable outcomes
and disregard those outcomes that are not likely. This however assumes that we have adequate amounts of insight
or diagnostic information to make our predictions reasonably accurate. It might be better to say that we spend less
time thinking about the outcomes that are not likely. When a particular outcome has a significantly low probability,
we might spend significantly less time thinking about it. This way, we can spend our time thinking about realistic ways to
accomplish our objectives, and therefore increase our chances of doing so.
We might wish to examine unlikely alternatives, for example, if we are performing analysis or are
in the middle or a correspondence chess game where we have the luxury of lots of time. There is a small probability that what
seems to be (initially) an unlikely course of action will somehow turn out in our favor, and in order to be thorough we might
wish to examine all courses of action to a certain depth of analysis.
p.187"A basic human tendency is to focus on the most probable outcome from among several that could
occur. This approach to decision making suggests that all except the most probable future be discarded... this criterion works
well when the most probable future has a significantly high probability so as to partially dominate."
