3 of 3 people found the following review helpful:
5.0 out of 5 stars
A Good Introduction to Simulation, July 23, 2011
This review is from: Introduction to Computational Science: Modeling and Simulation for the Sciences (Hardcover)
I enjoyed reading this book very much.
I found the book very useful in its conceptualization of simulation as a new form of synthesis for acquiring knowledge and helping human being make decisions. Simulation is a form of communication in that empirically-based models could be used to view the on-going processes that are cognitively beyond the capacity of human mind to untangle.
Furthermore, I found it admirable that the authors had instructed the readers in the art of model building using widely available simulation tools or even tools such as MS Excel that are not built specifically for the purpose of simulations.
I especially liked the tutorials with their wide selection of interesting material.
Regrettably, the subject matter of the book - simulation & (dynamic model building) - does not fit well within the traditional physics curriculum: Mechanics, Electromagnetism, and Quantum Mechanics. Numerical methods, including simulations, are not emphasized in such courses and normally one spends much of one's time studying well-known and solvable (in closed, analytical form) problems.
That does not mean that there is no room in physics for modeling and simulation: fractals, galaxy formation, dynamics of globular clusters, etc. are all areas that we are dependent on our numerical models and their fitness to observed phenomena to understand the processes of Nature. However, these are usually advanced topics not covered in undergraduate curricula.
And then the physicists tend to want to build their own tools rather that use COTS packages.
I think it is difficult to find a home in traditional university departments for a course on simulation based on this book. The fundamental reason, in my opinion, is that the personal computer as a scientific instrument is not valued or appreciated. And from that follows the lack of interest in simulations as venues for gaining scientific knowledge in Physics, in Chemistry & Biology. Conceivably a course in simulations might be of interest to engineers but then we would be leaving all those "soft"-science majors such as Ecology or Public Health behind. And those soft-science students are among some the people who could benefit the most from this book.
For example, the gene<->protein<->enzyme interactions, with their feedback loops and multiple pathways, are so complex that no human mind could expect to grasp all that goes on inside a cell. So Module 6.3 covering enzyme dynamics is absolutely on the right track from a scientific perspective; taking simulation out of the "hard" science world and into biology. Only through simulations and modeling are we going to develop a synthetic understanding of the cell in all its complexity.
I do not know if you have seen the book, "Historical Dynamics" by Peter Turchin in which he presents and discusses mathematical models of the evolution of agrarian states on the Eurasian land mass. His models are informed by empirical data collected from historical sources and do capture many aspects of historical reality. There is clearly a very important paradigm here at work but which is not as enunciated as I believe it should be; namely that simulations extend our scientific evidentiary-based knowledge into hitherto for dark realms of empirical experience.
Yet, Turchin did not use a simulation tool; he rolled his own and wrote much of his code in APL (A Programming Language) which is quite obscure. So readers must redo the models themselves. And his book is not about simulations, it is about what simulations tell us about history. I think that there are many fields of study in which the students could benefit from parts of this book; history, sociology, ecology, and natural resources comes to mind. There the exposition must be based heavily on using Commercial Off-the-Shelf (COTS) packages to tech the students how to build useful models.
So, for students in "soft" departments, the material in modules 5.2, 5.3, 5.4 could be skipped. In fact, even for students of physics and engineering, the understanding of the basis of the numerical simulations is not as important as learning how to build a system and observing how it behaves. On the other hand, the module 2, in my opinion, is important for all students to understand and to master so that they may interpret the results of their simulations correctly.
In my own case, I would love to be able to use a COTS package in which I could put galaxies - using a visual palette tool - on a 3-dimensional grid and observe their evolution in time as I changed the metric of space-time and/or the equation of state of the matter field.
Or consider the equations of stellar evolution, one would love to be able to run them again and again by changing parameters of these models knowing that the fundamental equations and their integration were worked out 70 years ago. As it is today, there are no such user-friendly generic approaches available to students or researchers, all such things must be painfully hand-crafted almost from scratch, barring some software libraries.
On a few occasions that I imagined writing such a book myself, I realized how difficult it was to do justice to the breadth and depth of the field of simulations from its hard-core physical scieces and engineering to ecology and wild-life management in a single book. While I might quibble with the inclusion of this or that topic or technique, I really cannot come up with a better design. There is an enormous amount of material here that may or may not be of interest to all audiences but there is a lot of material that is of interest to special audiences with focus on this or that scientific field.
The authors, if I understand them correctly, are positing that simulations (Computational Sciences) are a new way of knowledge discovery. In this they are right, in my opinion and are in the company of such luminaries as Dr. Steven Wolfram of Mathematic fame. But the problem is that "Introduction to Simulations" used to be taught in Industrial Engineering departments and then moved to places like the RAND Corporation and then the Pentagon. It is not viewed as a subject worthy of study in its own right (although such simulations as Halo or the World of War Craft sell millions of copies and make a few people wuite wealthy).
Which brings me to another topic related to simulations and to this book; namely computer games. Computer games are also simulations but with the caveat that they do not follow the Laws of Physics, Chemistry, and Biology and therefore are on somewhat of a tangent to "scientific" simulations such as those covered in this book. Gamers perform simulations, cosmologists do simulations, climatologists do simulations, agronomists do simulations, and many others but there is not set curriculum, no defined or unified approach and indeed no place to go to get an introduction to the art and science of simulation in a typical undergraduate college. This book is an attempt at just that.
I think non-science majors in Liberal Arts colleges could benefit also from a simulation course based on this book. Some colleges have a unified natural science department without the traditional divisions among Physics, Chemistry, Mathematics, and Biology. In such an environment, a course based on this book may offer that the non-science majors - a la Turchin's approach - the prospect of gaining confidence in building quantitative models of Reality and making inferences about the world on basis of such models.
This is a good book and a path-breaking book and I hope it finds the traction that it so righty deserves.
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
