SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

July 17th, 2006, 11:14 am

Below is a Table of Contents for a new book. Anyone who is interested can give input. We cannot promise to include everyhing but we will try to respond (at least) to all FAQs.For example, maybe there are topics that you would like to see that are not included in the TOC.Daniel J. DuffyDr. Joerg Kienitz///////////////////////////////////////////////////////The Project“Efficient and Robust Monte Carlo Methods in Financial EngineeringDesign and Implementation in C++”What is this Book?This book is about constructing, designing and implementing customizable software frameworks in C++. These frameworks realize functionality for the Monte Carlo method with a view to pricing, hedging (and calibrating) one-factor and n-factor option pricing problems. We apply a number of generic frameworks (from Duffy 2004, Domain Architectures) to allow is to create a framework that can be used ‘as is’ but that also can be used by QF people to suit their own needs. The architecture consists of a number of building blocks or components that we assemble to produce a working system.The first MC prototype is ready and the full architecture has been done for FDM. We apply this latter knowledge to MC in this project (book).What this book is notFirst, this book is not an introduction to the theory of the Monte Carlo method. There are several good books on this topic and we refer to them whenever possible (Glasserman, Jäckel). Second, we assume that the reader has some knowledge of the C++ language).Structure of this BookThe book has three main threads. First, we discuss the software architecture that is needed. Second, we apply it to one-factor models and finally we extend and apply it to challenging n-factor problems. In the last two parts we compare and contrast MC with FDM.Part 1: FundamentalsChapter 1 Analysis of Monte Carlo MethodsChapter 2 Designing Monte Carlo MethodsChapter 3 Building a Monte Carlo EngineChapter 4 Construction of Building Blocks for Monte CarloChapter 5 Integration and Software ArchitectureChapter 6 “Hello World” ApplicationPart 2: First ApplicationsChapter 7 Introduction to Path-dependent OptionsChapter 8 Asian OptionsChapter 9 Advanced Path-dependent ApplicationsChapter 10 One-factor Barrier OptionsChapter 11 Comparisons with the Finite Difference MethodPart 3: Advanced ApplicationsChapter 12 Introduction to N-Factor ModelsChapter 13 Correlation OptionsChapter 14 Stochastic VolatilityChapter 15 ‘Big Baskets’ (n ~ 50)Chapter 16 Comparisons with the Finite Difference MethodChapter 17 Early Exercise FeaturesAppendices1. An Introduction to the Finite Difference Method (FDM)2. Unified Modeling Language and Design Patterns3. Numerical Linear Algebra and Generic Data Structures4. Optimization Techniques5. An Introduction to Parallel Processing
Last edited by Cuchulainn on July 16th, 2006, 10:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
grafixel
Posts: 115
Joined: September 12th, 2002, 1:10 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 18th, 2007, 7:22 pm

Sounds great. Any plans when it should be finished?
 
User avatar
Collector
Posts: 4722
Joined: August 21st, 2001, 12:37 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 18th, 2007, 7:34 pm

Another book from you , what are you smoking to be so productive ? Or is it the Judo discipline transformed into book writing ? Congratulations!
Last edited by Collector on January 17th, 2007, 11:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 18th, 2007, 7:55 pm

QuoteOriginally posted by: grafixelSounds great. Any plans when it should be finished?Dr. Kienitz and your truly hope to be finished by July (manuscript). It will be a C++ customisable framework. It supports n factors.We use the classes and design patterns/generics. This helps.For the record, we do assume C++ and DP knowledge. There will be little overlap with (my) other C++ books.
Last edited by Cuchulainn on January 17th, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 18th, 2007, 7:59 pm

QuoteOriginally posted by: CollectorAnother book from you , what are you smoking to be so productive ? Or is it the Judo discipline transformed into book writing ? Congratulations!Thanks.Just lots of cups of Irish tea. And Joerg - co-author - of course, he is doing a lot. It's a real joint effort.Judo does indeed teach one some useful things. One is: pay attention to the fundamentals and continuous improvement. (also injuries sometimes )
Last edited by Cuchulainn on January 17th, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 22nd, 2007, 3:40 pm

QuoteOriginally posted by: grafixelAny plans when it should be finished?I think it is worth remarking at this stage is that we have the essentials of the OO/Generic framework in place with the necessary 'hooks' in place for the differnt kinds of derivatives, payoff and the other customisable classes such as RNG, output and so on. There is still much to do but the combination of the different styles that our programming language offers means that we can get a good performance on the one hand and functionality on the other hand. And the architecture consists of orthogonal components which will hopefully make the transition to a parallel implementation doable.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
ppauper
Posts: 70239
Joined: November 15th, 2001, 1:29 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 22nd, 2007, 8:49 pm

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: Collector what are you smoking to be so productive ? It's a real joint effortthat answer's collector's question !
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 24th, 2007, 7:11 pm

QuoteOriginally posted by: ppauperQuoteOriginally posted by: CuchulainnQuoteOriginally posted by: Collector what are you smoking to be so productive ? It's a real joint effortthat answer's collector's question !Wrong.Collector knows: we call it kiai
Last edited by Cuchulainn on January 23rd, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Collector
Posts: 4722
Joined: August 21st, 2001, 12:37 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 25th, 2007, 3:51 am

Yes kiai (now we are talking) And as students of judo we know how to fall . Life is a series of falls (not continuous time Brownian Motion) so if you know how to fall softly you can quickly raise again. But even knowing how to fall you must expect a broken arm from time to time, use the arms to protect your head, you can write another book without a arm, but not without a head.
Last edited by Collector on January 24th, 2007, 11:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 25th, 2007, 7:03 am

My teacher was Willem Ruska, who got 2 gold medals judo in Munich 1972, then one learns to fall well. here
Last edited by Cuchulainn on January 24th, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 25th, 2007, 8:26 pm

Here is a (possibly) subjective summary of why C++ is suitable for MC frameworks and derivatives applications.Comments and critique welcome.here
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 27th, 2007, 9:11 am

Someone had some questions concerning MC and which language to use...Basically, I should be doing it in C# instead of C++. Fair enough. We are happy with each language but I feel that quants prefer C++?What are the opinions on this? Edit: porting the code to C# is not so difficult, it could be a nice (student) project?
Last edited by Cuchulainn on January 26th, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
dibble
Posts: 727
Joined: October 2nd, 2006, 5:19 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 27th, 2007, 4:29 pm

Linux, Mac os, Windows => C++ edit: Of course the same equation would give you java, but since the question did not mention java, so this would be the wrong answer
Last edited by dibble on January 26th, 2007, 11:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Topic Author
Posts: 62887
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 28th, 2007, 10:17 am

QuoteOriginally posted by: dibbleLinux, Mac os, Windows => C++ edit: Of course the same equation would give you java, but since the question did not mention java, so this would be the wrong answerC++ is ubiquitous and portable. We need the code in Linux as well because we wish to do some n-factor MC modelling in MPI. So it is a fact that Linux is most suitable, fine by me. You could the front end in friendly Windows and have an ACE socket to the Linux cruncher.No experience of MACOS. I think java is not the most appropriate metaphor here. But C#?
Last edited by Cuchulainn on January 27th, 2007, 11:00 pm, edited 1 time in total.
Step over the gap, not into it. Watch the space between platform and train.
http://www.datasimfinancial.com
http://www.datasim.nl
 
User avatar
grafixel
Posts: 115
Joined: September 12th, 2002, 1:10 pm

Forthcoming Book: Monte Carlo and C++ (Kienitz/Duffy)

January 28th, 2007, 11:16 pm

MC is compatational intensive stuff, i never would even think about using more highlevel language for that (unless you have pretty much time to wait for the results). C++ is already a compromise compared to C (reading other threads even fortran seemingly still plays a role). Paralell computing plays a role here - managing this might introduce other aspects. i am no expert rgd this however
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On