Mathematical modeling and optimization. An essay for the design of computer-based modeling tools (Q1809328)

My colleagues and I once created a mathematical model to explore the communication behavior of a mesh based network [\textit{Z.-S. Shen}, \textit{P. Drexel} and \textit{L. Urbach}, Math. Comput. Modelling 18, No. 12, 33-48 (1993; Zbl 0805.68013)], by using such mathematical and logic notations as first-order logic, linear algebraic, probabilistic, and various combinatorial expressions. Since these declarative statements can't be executed directly, later on, we had to ask a student to convert this mathematical model into an executable program in \(C\). After these many years, I still can remember the pain it took the student to understand what I meant by those ``weird'' expressions and the effort it took him to do the conversion. Apart from the quality of such a model, I wish at that time there were a better way for its construction and execution. Moreover, although I am not a professional modeler, I certainly hope to have some general understanding of the whole modeling process. Now, when I am reading this wonderful book by Hürlimann, I definitely wish it were available when I wrote that model. It has to be pointed out first that this book is not intended to hand over any magic solution to the users to build up their models. It is rather intended to help the modeling tool designers understand the whole modeling process better in order to develop more user friendly tools for the modeling practitioners. It is like a book written for the car designers so that we can have better cars to drive. As a consequence, it shows what could be a modeling language, and why, but not how to develop a model for a specific problem. This book consists of three parts: General concepts (Chapters 1-4), a discussion on modeling management systems (Chapters 5-7), and a proposal for such a system (Chapters 8-10). More specifically, after an introductory chapter, Chapter 2 defines the notion of modeling, particularly, mathematical modeling, and some of the related concepts. Many examples, old and new, simple and complicated, artistic and scientific, as well as some theoretical ``stuff'', are used to demonstrate and justify both the ends and the means of the modeling process, referred to as our ``power of abstraction''. Chapter 3 presents a modeling life cycle, beginning at specification, and ending at report writing, that summarizes various problems a modeler has to tackle. Chapter 4 contains an overview of different types and paradigms of mathematical models, such as analytical vs. simulation, linear vs. nonlinear, symbolic vs. numerical, etc.. It ends with a discussion of modeling uncertainty, by presenting various techniques and their inclusion in the respective model building process, which are well exemplified. Chapter 5 argues for, discusses the status quo of, and explores many issues related to, a general, uniform, modeling management system, i.e., a set of tools ``to provide computer-based assistance to the decision maker for creating, solving, manipulating, and integrating a variety of models'' [\textit{B. Shetty}, \textit{H. K. Bharagava} and \textit{R. Krishnan} (eds.), Model management in operations research, Annals of Operations Research 38, 1-4 (1992), Preface]. Chapter 6 presents a critical review of several frequently used approaches for modeling, such as spreadsheet, database, graphic modeling, CLP, algebraic languages, as well as some of the proposed general modeling frameworks, including a structured modeling, an embedded language approach, and a multi-view architecture. Chapter 7, ``the core chapter'', presents the author's own approach to the development of such a system, the LPL language [\((*)\)] (currently known as Logical Programming Language), with its implementation being discussed in the following Chapters 8 and 9. This language allows the modeler to specify the structure of the problem declaratively, while being able to express a solution algorithm procedurally. It also comes with an environment to support various steps of the aforementioned modeling life cycle. In the concluding Chapter 10, more than a dozen problems taken from various application areas are discussed, and the corresponding LPL codes given. This book ends up with almost twelve pages of references. I find the attached LPL WEB site [\((*)\)] [\url{http://diufpc03.unifr.ch/lpl/lpl.html}] not fancy, but quite useful. It providesa collection of LPL related papers and technical reports, including manuals, besides a free package of the LPL system. It also has an interactive feature. You can simply type in the LPL model for your problem, a built-in model solver will solve it, and display the results. Besides being rigorous, this book is quite readable. It is an ``essay'' after all. I definitely agree that besides stimulating modeling tool designers, this book can also be used as a textbook in a mathematical modeling course, perhaps accompanied with a LPL manual. Moreover, this book, particularly, the first part, maybe the second as well, is very informative, and should be interesting to more general audience. For example, now I know where to look for real life examples when I teach the modeling concepts in the beginning computer science courses.