Skip to main content
main-content

Über dieses Buch

People use the word strategy in a variety of different contexts. The term has connotations ranging from statesmanship to economic planning, and has become pervasive in the social sciences. We also talk about "problem solving strategies" and "corporate strategy" in a large business enterprise. The concept of strategy applies whenever a sequence of goal-oriented actions is based on large-scale and long-range planning. This monograph gives a systematic overview of the theory of strategies, a new area of enquiry developed over the past two decades by the author and his team. The projects described have clearly defined research objectives and are based on realistic assumptions about the environments in which the programming systems will work, and about the constraints and requirements they have to satisfy. Applications of the systems range over various aspects of air traffic control, automatic verification and validation of discrete-event simulation models, econometric model building, distributed planning systems for manufacturing, control of traffic lights, and others. The book is aimed at researchers, teachers and students in computer science, management science and certain areas of engineering. The reader should have some maturity in computer science and mathematics, and familiarity with the basic concepts of artificial intelligence.

Inhaltsverzeichnis

Frontmatter

1. Introduction

Abstract
In addition to to the millions of ‘mundane’ tasks in business, engineering and scientific calculations, computers have also been used in discovering, modifying and verifying scientific theories. Let us discuss briefly and informally what is meant by a scientific theory and how can it be computer-based. This area is probably the most controversial in Artificial Intelligence and is closely related to the philosophical problem of how machines can manifest creativity and originality. Further, I note that the epistemological relation between theory and program is held by some people as (i) a possibly complete identity, or (ii) the appropriate program is considered as the model of a theory, or (iii) the program is taken as a heuristic tool in theory formation but with no formal connection to theory. I am not a philosopher and it is not my intention to cover such questions in this book. In forming one’s opinion, the empirical observations are obviously strongly affected by attitude and perception. The following informal discussion should be regarded simply as a guide to orient the reader before embarking on the main topics.
Nicholas V. Findler

2. Basic Concepts of Strategies, Decision Making, and Planning

Abstract
People use the word strategy in a variety of different contexts. Its original meaning, “the art of the general” in ancient Greek, refers to the conduct of warfare1. The term has later assumed connotations ranging from statesmanship and management of national policy to diplomacy and economic planning. Ever since John von Neumann and Oskar Morgenstern [162] showed the similarity between game-like problems in economics, sociology, psychology and politics, the concept of strategy has become pervasive also in the social sciences. We also talk about “problem solving strategies”, “corporate strategy” in a large business enterprise, and so on — whenever a sequence of goal-oriented actions is based on large-scale and long-range planning.
Nicholas V. Findler

3. The Quasi-Optimizer (QO) System

Abstract
One usually distinguishes in science between descriptive and normative theories. The former tells us what an entity does under certain conditions and, possibly, why. Cognitive psychology, sociology and demography are examples of the disciplines that provide descriptive aspects of human behavior. In contrast, normative theories tell us what to do in order to accomplish certain goals. Various optimization techniques and methods of Operations Research yield normative theories of well-defined and mathematically formulable tasks.
Nicholas V. Findler

4. The Advice Taker/Inquirer (AT/I)

Abstract
The Advice Taker/Inquirer is a domain-independent program that is used to construct, fine-tune and oversee the operation of an expert system. It consists of two phases: a learning phase, during which a human expert teaches the system interactively about a domain in terms of principles, high-level examples, rules, and facts; and an operational phase, during which the program monitors the resulting expert system as it is applied to a domain of interest, and continually attempts to improve its performance by hypothesizing new rules and reorganizing existing knowledge.
Nicholas V. Findler

5. The Generalized Production Rule System (GPRS)

Abstract
In decision making, a fairly precise knowledge of the values of all relevant situational variables is necessary. We call open variables (OVs) those that can be observed and measured at any space and time point while the values of hidden variables (HVs) can be obtained only intermittently, at certain points of space and time. The following examples should clarify this issue:
  • Atmospheric conditions. The OVs are measured continuously at the Earth’s surface whereas high-altitude variables, such as stratospheric wind velocity and air temperature, are HVs and can be observed only when, for example, balloon-borne instruments and a radio transmitter are sent up to the “right” altitude and location.
  • Material testing. Inexpensive and non-destructive testing of some material or product can be performed at any desired time but costly or destructive tests are carried out only sparingly.
  • Oil and mineral exploration. The evaluation of satellite photographs, seismic experiments, geological surface studies and deep-drill work can produce OVs or HVs, depending on the associated factors of cost and difficulty.
  • Earthquake prediction. In addition to seismographical data, a number of different OVs have been proposed and used that are likely to be correlated with the near-future occurrence of earthquakes. The HVs can be, for example, the location of the epicenter and the intensity of an earthquake to come.
  • The training of specialists. The (controlled) OVs may be the length and intensity of training or the frequency of given exercises, while the level of actual performance under certain rare conditions represents an HV.
Nicholas V. Findler

6. Distributed Planning and Problem Solving Systems (DPPSS)

Abstract
In continuing the ideas presented in Chapter 2, it will helpful if we consider plans as representations of proposed courses of action and of the agents to carry out the actions. Human or machine decision makers have to reason about and evaluate plans before and during plan execution.
Nicholas V. Findler

7. Causal Modelling Systems (CMS and NEXUS)

Abstract
At the risk of oversimplifying concepts and terms that are important to philosophers, the elements of processes involved in scientific inquiry — observational procedures, sequences of argument, methods of representation and computation, concerns about validation — basically aim at discovering causal relations between patterns of empirical phenomena. On the basis of these causal relations, science has the goal of constructing concise and systematically organized theories that describe and explain the world of nature. Technology then makes use of the results of the scientific enterprise to build reliable, effective and economically justifiable products to serve given needs and objectives.
Nicholas V. Findler

8. The Predictive Man-Machine Environment (PMME)

Abstract
In the previous chapters, we have described several large-scale programming systems which aim at combining the strength and flexibility of humans and computers in decision making, problem solving and planning tasks. It seemed desirable to bring these systems together so that they all can contribute to the solution of some difficult problems. The logical scheme of this integrated programming environment is shown in Fig. 8.1.1.
Nicholas V. Findler

9. Overall Summary

Abstract
We have discussed in this book a number of projects, each of which relates to some aspects of decision making strategies — strategies, for short. We have considered a strategy both as a guide to making plans and solving problems, and a mechanism that observes a part of its environment, regarded as relevant by the strategy, evaluates it and prescribes a certain response to it.
Nicholas V. Findler

Backmatter

Weitere Informationen