Guohua Bai (Ph.D. Guest Professor)

 

2010-02-28

My reading notes

 

 

Notes of Systems (cybernetic) Theory Study

Part 1 ; Part2; Part 3; Part 4


PART FOUR


Nadler, G. (1985)

Systems Methodology and design.

IEEE transactions on systems, man, and cybernetics, Vol. SMC-6(15): pp 685-697.

Concept of design:

System methodology and design incorporates various disciplines, philosophies, schools of thinking, approaches and actural practices. Discussing SMD cannot be limited to a single perspectives, or only to one level of abstraction. Therefor several levels of abstraction or perspectives are presented, followed by an attempt to define the bridges between these ways of thinking.

Overall factors for understanding the world around us, usually presented in dichotomous terms in the literature, need to be considered before moving on to design.

1) Reductionism and expansionism:

This belief says that design can be reduced, decomposed, or disassembled to simple "indivisible" parts. Design of the whole is the sum of the designs of its "independent" parts. Reductionism is based on the notion of a "mechanism," which relies solely on the relation of cause and effect, thus presumably allowing subparts to be integrated together to form the whole design. Expansionism: This view maintain that systems are part of larger systems, and therefore designs are part of larger designs. But the emphasis of design here is redirected from the ultimate elements to a whole design with interrelated parts. The emphasis is on the functional entity that contains and structures parts and elements.

2) Inductive & deductive:

An inductive process seeks generalization through to unify a series facts A deductive process seeks a situation-specific solution from and based within the premises of human organizational needs, and reasonably available generalizations or theories. (Synthesis) Deductive process combined with expansionism leads to a "system approach". It is this approach that will be shown to be the most appropriate for design.

3) Systems vs. Conventional:

a) Advantage of the systems approach:

The whole is more than the sum of its parts The part is more than a fraction of the whole

___ Aristotle (The composition Law)

The solution produced by conventional thinking is usually quite limited because the designer often gets involved in doing well what may not need doing at all, instead of focusing on the guidlines of purposes and what ought to be done. It forces designers to adopt a technique orientation, that is, finding where a technique can be applied or looking for the technique or equation rather than determing what problems need solving or purposes need achiving.

Designer as Black Boxes: Output effectiveness of designer depands on input, feedback, objects, environment.

Designer as glass boxes: Variables, objects, criteria are fixed in advance. Analysis is completed before solutions are sought.

Designers as self-organizing system: (nothing new in this)

Types of problems designers face:

a) improve an existing system or product

b) Diagnose or remedy some trouble or find a cause.

c) develop a new system or product.

d) develop a new use for an existing system or product.

The conventional methodology:

a) Gather Information and facts b) Formulate the problem c) Explore alternatives d) Select the solution e) Detail the solution f) Implement the solution.

Systems methodology:

a) determine the needed purposes in its hierarchy of purposes. b) generate purposeful alternative ideas. c) develop feasible idea solution target. d) develop and detail solutions. e) install proposed design and follow up.

C: Common Traits Found in Using Systems methdology and design:

1). Hierarchical nature of Design: For want of a nail a shoe is lost (found)

For want of a shoe horse is lost (found)

For want of a horse a rider is lost (found)

-- George Herbert

Reviewing hierarchies does not mean isolating each level. The relations and bonds between levels of hierarchy are essential.

2). Functional nature of design:

Design has the function of developing purposeful systems. Everything in the designed system is considered and evaluated as related to the purpose(s) of its larger systems.

3). Process of design:

Design is a process of decision making. decision making is necessary in each step of design without which progression is not feasible. An important reason for the system design process is to improve decision making with repect to the objectives of the application area. To make a decision is to select a course of action from several.

4). Iterative nature of design:

The iterative concept has its roots in control theory and cybernetics. Iteration is continued untill a percived equilibrium point is reached. The future system will exist around this equilibrium or will try to move toward this equilibrium untill a new equilibrium is found on the horizon.

5) Treatment of optimization:

local optimization and global optimization.

D. Omissions in the literature-unanswered questions:

1) Design for the future:

Nothing is permanent except change.

-- Heraclites

Concrete suggestions on making the changes in inputs and objectives as part of the design process is lacking in many methodologies.

2) Creating Alternatives:

Black box method is one of the most entirely applied method.

3) Determine the hierarchy of needs:

4) Decisionmaking:

5) Selecting information

6) problems of implementation

7) Problems of rejecting a "good" alternative:

"The sublime and the ridiculous are often so nearly related that it is difficult to class them separately. One step above the sublime makes the ridiculous, one more step makes the sublime again" -- Thomas Paine

8) Which methodology to use and when?:

Each proposed methodology claims to be the most appropriate for all the designer´s problems. Although such claims are often supported by some kind of ligical reasoning, empirical studies are needed to evaluate the whole and the parts of various methodologies, and to produce guidlines about which methodologies are best suited for the purpose at hand.

9) Integration of science:

For either you know what you are looking for and then there is no problem Or you don´t know and then you cannot expect to find anything.

-- Plato

When it comes to engineering design, the entire fields of behavioral science, cognitive science, decision analysis, group technology, psychology, etc., are almot untouched. The philosophical nature of information (e.g., information is not a brick which can be tossed from one person to another with exact same meaning) needs wide discussion.

III. The Need For a Unified Systems Design Concept for Engineering

"Complexity has a way of feeding on itself. As the complexity of a problem cause many minds to be applied to it, a new complexity arises, that of aharing among these many minds the products of each other´s thoughts and actions and this in turn leads to another complexity, that of interrelating and distilling group effort and bringing them into a format where they assume utility."

A: Several Guidlines are available for developing and promotig SMD.

( This list could be used as criteria of Feedback-learning design methodology)

1) It should be tailored to a very wide area of problems

2) Its parts (steps, formats,etc.) should be largely context free, in the sense that they are not addressed to any apecific academic discipline or content area of application; all teams have access to such structure.

3) Its parts should be useful in specific components of problem solving, but not so gross and general so as to defy the construction of meaningful examples of their use.

4) Communicability across disciplines should be a characteristic of the SMD parts in order not to constrain the utilization to persons with specific kinds of disciplinary background.

5) Full development of a part should not be a requirement for inclusion in SMD. The part should be recognized as needed.

Systems basis in philosophy:

A whole branch of philosophy(cosmology) deals with the universe as an orderly system and looks towards an ideal of single cosmology unifying the many sciences. Systems basis in psychology:

* Human need: This can be considered as the motivation for all forms of human endeavor. Moslow described a hierarchy of needs of human being.

* Creativity Behavior

* Group problem solving

Systems basis in mathematics:

modeling quantitative relational forms, dynamic analysis, optimization, cybernectics.

The following characteristics distinguish the systems approach in respects to what are required:

a) Deal with problems of organized complexity;

b) emphasizes broad considerations and many interrelations;

c) requires far-reaching, often controversial value judgements;

d) incorporates careful planning;

e) employs tools specially deigned to deal with complexity; and

f) requires outstanding management of people and resources.

The factors emerge a need for a merger between engineering pragramatic bases and academic bases in systems approach:

a) The conduct of systems engineering on a strictly programatic basis tends to force the discipline too far in the direction of routine work where the effort, rather than the results, tends to be an end in itself.

b) Both bases may gain strength from the merger.

c) A merger would help define the roles needed to conduct the systems approach.

d) Merger of what should be done and how to accoplish it.

Benfits of the SMD:

a) Increase the probability of working on the right problem ( not developing a good design for what ought not to exist at all).

b) Holistic view of design. optimization as an improvement of the whole rather than a part.

c) It leads the design as a dynamic process rather than a static state, so learning is absolut needed all the time.

d) People with different values, expertise, experiences, baises, etc. are equiped to deal with and take advantage of.

e) It increases the chances of getting managerial and stakeholder approval for installing the solution,

since they are involved immediately and continuously at eah step of design where they participate in the decision process.

f) Design is recognized as a continuous process extending beyond the initial installation, according to the dynamic nature of the real world.

Dash, D.P. (1994)

Systems Dynamics: Changing perspectives.

System Practice, Vol.7, No.1, pp87-98

The systems approach goes on to discover that every worldview is terribly restricted

-- C.W. Churchman

The origin of ystems dynamics (SD) was started at the Massachusetts Institute of Technology (MIT), Cambridge, around 1958, under the pioneering leadership of Jay W. Forrester, who found that the concepts and ideas of control theory and control engineering could be made to apply to management, and more generally to socioeconomic problems, in the same sprit with which they applied to design physical systems (Industrial Dynamics, 1961). The SD paradigm is rooted in the premise that system behavior depends upon system structure.

System: those aggregates of physical and abstract entities as systems which are distinguishable from their surrounding envirnment as purposive whole and which exhibt dynamic behavior.

The basic strategy of SD is modeling, which is operationalized theough a procedure akin to the following.

1) define the system to be studied.

2) define what improvement is to be sought.

3) represent system structure through a structural model.

4) translate the model into mathematically computable relations.

5) test and validate the model.

6) evaluate alternative policy options.

Critique of SD:

1) dificult to define the system because of system interests are so intimately interwined with their surrounding environment that the model tend to grow in size substantially before a meaningfull representation is achived, and difficul to identify which elements influence which others and in what way.

2) difficult to validate and test the model because unvailability of reliable data

3)ontological valid ? philosophical question.

4)Subjective bias, the model are actually influenced by the values, beliefs, and goal of the modelers.

5) too hard: This is stated "man is no mere biocybernetic device" (Meadows, 1974).

6) utopian: Experience in SD practice has shown that sometimes the recommendations arising out SD modeling exercises are not implementable due to reasons of politics anfd culture. In a decision situation, what appears to be logically best alternative does not seem to be clearly the most persuasive for the problem owners to adopt.

Some conceptual shift of SD after those critiques:

1) "process" image of Models. Modeling is now seen as a continuing companion to, and tool for, improvement of judgment and human decision making.

2) Enriching policy Debates. Forrester has emphasized that description of the real world is not the objective of SD models but models should become part of a more peruasive communication process that interact with people's mental models, creates new insights, and unifies knowledge.

3) Model simplification: The elaborate and complicated models developed do far have been found to be intellectually inaccessible to problem owners. Therefore those models tend to remain in reports and are never really used.

4) Models for insight. Models are now judged by how much more "insight" they generate on the nature of the system being modeled, instead of how "valid" they are (Forrester, 1987).

5) Qualitive analysis

6) SD plus. Combine other relevant theories and methods.

7) relativim

SD today

P.N., Senge (1990) emphasizes feedback thinking in organization in order to enhance their learning ability. His focus is not as much on the controversies regarding SD paradigm, as it is on a more widespread use of the classical SD concept.

Hutchins, Edwin (1990)

Organizing Work by Adaption

Organization Science, Vol 2, No.1, pp14-39.

I will examine the response of a work group to a change in its information environment. I will argue that several important aspects of a new organization are achieved not by conscious reflection but by local adaptions. The solution reached ..... is a product of adaption rather than of deign.

A case study of navigating large ships (an incident when the ship's propulsion failed unexpectedly).

Dahlbom Bo & Mathiassen Lars (1991)

Struggling with Quality-The Philosophy of Developing Computer Systems.

Report 4, Gothenburg studies in information systems.

View points are expressions of different interests and inherent contradictions, e.g., between systems developers and users or between managers and employees. The dialectical systems approach is based on the Romantic idea that the world is always changing, and that we cannot understand it unless we understand what change is and why it takes place. The claim of the dialectical approach is that we must think it in terms of contradictions in order to understand, explain and control change. In the dialectical systems approach, contradictions appear not only in our thinking, they are in the world itself. Reality is assumed to be a totality of related contradictions, its most dominant feature being change. In any given situation we face a network of related and dynamically changing contradictions. Some of these are more dominant than others, and in each of them one of the two opposites is more dominant than the other. As the situation changes, other contradictions will become more dominant, and the opposites of each contradiction will be differently balanced.

The special nature of human activity systems means that systems studies concerned with them are always multi-valued, with many relevant and often conflicting values to be explored. The outcome is never an optimal solution to a problem, it is rather a learning which leads to a decision to take certain actions, in the knowledge that this will in general lead not to 'the problem' being now 'solved' but to a new situation in which the whole process can be again. (P.Checkland, sytems thinking, systems practice, p213)

The fact is that no contradictory aspect can exist in isolation. Without its opposite aspect, each loses the condition for its existence. ....It is so with all opposites; in given conditions, on the one hand they are opposed to each other, and on the other they are interconnected, interpenetrating, interpermeating and interdependent, and this character is described as identity (Mao,1937, p338)

There are many contradictions in the proces of development of a complex thing, and one of them is necessarily the principal contradiction whose existence and development determine or influence the existence and development of the other contradictions (Mao, 1937, p331).

The universality or absoluteness of contradiction has a two fold meaning. One is that contradiction exists in the process of development of all things, and the other is that in the process of development of each thing a movement of opposites exist from beginning to end. (p316).

... in every form of motion, each process of development which is real (and not imaginary) is qualitatively different. Our study must emphasize and start from this point (p320-321)

Ackoff, Russel L. (1995)

'Whole-ing' the parts and righting the wrongs

Systems Research Vol.12 No.1, pp 43-46

Architects provide a paradigm of systems practice that should be emulated. They seldom improve parts at a cost to the whole. They 'whole' the parts unconsciously, as an inherent aspect of their design process.

Learning takes place when a mistake is identified, its producers are identified, and it is corrected.

It is better to do the right thing wrong that the wrong thing right; the former leads to learning; the later to reinforcement of error.

To do the wrong thing right is to do it efficiently but not effectively. Effectiveness is evaluated efficiency. Therefore, effectiveness is obtained only when the right thing is done right.

"Each level of integration resolves the contradictions of the levels below and so removes the errors that were most serious there. Each level brings about an improvement of judgment. Each level exhibits more truth through higher integration of the facts. There is much more truth in Ptolemy than in Anaximenes, ore in Kepler than in Ptolemy, more in Newton than in Kepler. It appear that the criteria of truth are precisely the features of the organic whole-inclusiveness, determination, and organicity...."

Return to Reading Notes