** PLEASE DESCRIBE THIS IMAGE **

University College Dublin


Derek O'Connor

Retired Jan 2000

This page was last updated on October 14, 2012

derekroconnor[AT]eircom[DOT]net


Past Courses


John F. Muth (1930-2005)

Jack Muth was my doctoral supervisor at Indiana University, 1973-77. He is best known for inventing Rational Expectations. Here is an obituary by Ike Brannon at the Cato Institute :

Remembering the Man Behind Rational Expectations

Unlike Brannon, I did not find him shy or socially awkward, but he could be very sarcastic. I still cherish his comment at a student party when he received his 10th drink : 'I asked for vodka, not gin.'


Notes & Essays

The following notes etc., are on my Scribd page here http://www.scribd.com/derekroconnor4276

  1. The Birthday Paradox and Random Number Generation

  2. RPGLab: A Matlab package for Random Permutation Generation

  3. Matrix Chain Multiplication

  4. Two Simple Statistical Calculations and ClimateGate

  5. Floating Point Arithmetic -- DAMC Talk, 2005

  6. Matlab's Floating Point System

  7. RPGLab Matlab functions

  8. Alglat - A Minimum Change Algorithm

  9. Stochastic PERT Networks

 

Interesting Papers

  1. Kalman -- A New Approach to Linear Filtering, etc. (1960)

  2. Moler & Morrison -- Pythagorean Sums

  3. URAND generator from Forsythe, Malcolm, & Moler

 

Stochastic PERT Networks

The Stochastic PERT Network (SPN) Problem is #P Complete and so exact solutions are possible only for relatively small networks. Below are two related papers on this subject.

1. Exact and Approximate Distributions of Stochastic PERT Networks. (Pert2006.pdf 240KB)

Abstract. This paper presents a definition and a simple characterization of Conditioning Activities which, along with a simple theorem, leads to an algorithm that can be used to calculate exact or approximate (Monte Carlo) distributions for the completion time of Stochastic PERT Networks.

The time complexity, for networks of \(n\) activities, is exponential for exact distributions and \(O(n^2N)\) for Monte Carlo approximations with sample size \(N\). Computational tests are reported for 4 networks varying in size from 10 to 40 activities.

2. Minimum-Change Algorithm for Generating Lattice Points. (Alglat2006.pdf 89KB)

Abstract. This paper presents a simple minimum-change algorithm to generate all lattice points in a given volume. The running time of the algorithm is \(O(N)\), where \(N\) is the number of points to be generated.

 

Dublin Area Mathematics Colloquium Talk Floating Point Arithmetic, March 4, 2005.

 

Kalman Filters

R.E. Kalman : "A New Approach to Linear Filtering and Prediction Problems", Transactions of the ASME - Journal of Basic Engineering, 82 (Series D) , pages 35--45, 1960. This is Kalman's classic 1960 paper, transcribed from the original into PDF by John Lukesh. Kalman1960--Annotated.pdf (250KB)

 

MMS Dissertation

Kelly, D.J. & ONeill, G.M : "The Minimum Cost Flow Problem and The Network Simplex Solution Method", M.Mangt.Sc. Dissertation, University College, Dublin, 1991. MMS-91.pdf (394KB)

 


Shortest Path Problems

This is a small set of real road networks stored in Matlab's sparse matrix format : SPathProbs

Software

Here is some information on free or cheap alternatives to Matlab and Maple : software

Here are some benchmarks of Matlab and clones along with multicore tests : benchmarks


Lincoln Steffens Memorial Prize : The nominees may be viewed at Steffens2006.


Directions and Map from Airport to Donard Co. Wicklow (750K PDF)

Curriculum Vitae


Locations of visitors to this page                             Visits since Sept 2004     ** PLEASE DESCRIBE THIS IMAGE **