Catálogo de publicaciones - libros

This chapter introduces the new systems development platform presented in this book. This new development methodology for management information systems (MISs) is based on a formal model-theoretic structure derived from the systems concepts of general systems theory (GST). The model is represented in set theory and implemented in a fourth-generation programming language, extProlog, by automated system generation. The extProlog language is an extension of standard Prolog that allows for the implementation of an MIS. As discussed in Section 1.2, the new methodology provides a platform for the development of both transaction processing systems and problemsolving systems as the two principal components of an MIS.

To realize an information system by the model theory approach, a user model in set theory as introduced in Chapter 1 must be described in a form that can be accepted by a computer. For example, the symbol “∉” cannot be accepted by a computer, and must be replaced by an appropriate term, in this case “notmember”. The system-defined predicates and functions are used to describe the predicates and functions necessary for the user model with greatest efficiency. Set theory created in this way is described here as computer-acceptable set theory, and is the focus of this chapter.

This chapter introduces extended Prolog (extProlog). Although primarily used as an implementation language for the model theory approach in a hidden manner similar to a machine language, knowledge of this language is also helpful for understanding the model theory approach, for the following reasons:

This chapter outlines the process of problem-solving system development by the model theory approach as a preliminary for Chapter 5. Specifically, development of extSLV is the target. A problem classification system is introduced for development. Throughout this book, extSLV development is carried out based on this classification scheme.

This chapter presents detailed explanations of the design stages of Fig. 4.6 and theoretical results from the user models and standardized goal-seekers featured in Fig. 4.5. The theoretical results can be omitted if readers are interested in practical aspects of the model theory approach. There are eight cases of user models according to the classification of Section 4.3. This chapter investigates two extreme cases: the E-C-C case and the I-O-O case. The other cases can be considered as combinations of these two extreme cases.

This chapter examines the case E-C-C of the problem classification presented in Table 4.1. The famous traveling salesman problem is used as a typical example of the case. We develop extSLV for the problem following the development procedure of Chapter 4.

Chapters 7 and 8 discuss a control engineering problem. Control engineering problems are interesting for three reasons. First, although control is a type of management, because a target of control engineering is a continuous dynamic physical system, a control engineering problem is naturally different from a management problem. On the other hand, the methodology of this book has been developed and aimed at management problems. However, because it is based on the concepts of general systems theory (GST), it should be a general theory, which implies that our methodology must be applicable to control problems. It is interesting to see how our theory can work as a general theory in a field that is considered outside the scope of the original intention.

Chapter 7 introduced a simple dynamic system as a control engineering problem and investigated it as an E-O-C problem. This chapter formalizes an optimization problem for a general second-order dynamic system and examines it as an E-C-O problem. The goal is defined as a traditional quadratic problem. Because the goal of the optimization problem has parameters that can be modified by a user, the problem can also be considered as an E-O-O problem. This chapter introduces a constraint condition to investigate the significance of backtracking, which is the most basic characteristic for distinguishing the goal-seeker from conventional controllers.

This chapter discusses a cube root problem as a simple case of an I-C-C problem. As shown below, an extended solver (extSLV) for this problem can be easily developed when the model theory approach is applied because the problem is simple. It is, however, puzzling to see that we cannot find a nontrivial problem for the case of I-C-C, although there are many tough problems for its neighboring cases of E-O-C, I-O-C, E-C-O, and I-C-O. For example, the regulation problem of Chapter 7, an E-O-C problem, can become a tough problem if the number of bodies, which must be controlled, increases. The magic square problem of I-O-C is quite difficult if its size becomes large. The conventional dynamic optimization problem, in Chapter 8 falls into the E-C-O class. The famous knapsack problem described in Chapter 10 is an I-C-O problem.