Maximum Likelihood Estimation Using Parallel Computing: An Introduction to MPI

The computational difficulty of econometric problems has increased dramatically in recent years as econometricians examine more complicated models and utilize more sophisticated estimation techniques. Many problems in econometrics are `embarrassingly parallel' and can take advantage of parallel computing to reduce the wall clock time it takes to solve a problem. In this paper I demonstrate a method that can be used to solve a maximum likelihood problem using the MPI message passing library. The econometric problem is a simple multinomial logit model that does not require parallel computing but illustrates many of the problems one would confront when estimating more complicated models.

[1]  Anna Nagurney,et al.  A Splitting Equilibration Algorithm for the computation of large-scale constrained matrix problems: theoretical analysis and applications , 1992 .

[2]  Jack Dongarra,et al.  PVM: Parallel virtual machine: a users' guide and tutorial for networked parallel computing , 1995 .

[3]  Anthony Skjellum,et al.  Using MPI: portable parallel programming with the message-passing interface, 2nd Edition , 1999, Scientific and engineering computation series.

[4]  Ian T. Foster,et al.  Designing and building parallel programs - concepts and tools for parallel software engineering , 1995 .

[5]  Benjamin Ray Seyfarth,et al.  How to Build a Beowulf: A Guide to the Implementation and Application of PC Clusters , 2000, Scalable Comput. Pract. Exp..

[6]  Wilbur John Coleman.  Solving Nonlinear Dynamic Models on Parallel Computers , 1993 .

[7]  Michael Florian,et al.  Parallel Implementations of Primal and Dual Algorithms For Matrix Balancing , 1994 .

[8]  Hans M. Amman,et al.  Computational economics and econometrics , 1992 .

[9]  A. E. Fincham,et al.  Parallel Computation , 1999, Algorithms and Theory of Computation Handbook.

[10]  G.B. Williams,et al.  Use of parallel supercomputing to design magnetic resonance systems , 2000, IEEE Transactions on Applied Superconductivity.

[11]  Hojjat Adeli,et al.  High-Performance Computing for Large-Scale Analysis, Optimization, and Control , 2000 .

[12]  G. Amdhal,et al.  Validity of the single processor approach to achieving large scale computing capabilities , 1967, AFIPS '67 (Spring).

[13]  Robert B. Ross,et al.  Using MPI-2: Advanced Features of the Message Passing Interface , 2003, CLUSTER.

[14]  Robert A. Moffitt,et al.  Incentive Effects of the U.S. Welfare System: A Review , 1992 .

[15]  D. Rubin,et al.  Markov-Normal analysis of iterative simulations before their convergence , 1996 .

[16]  Gotz Uebe.  Computational techniques for Econometrics and Economic Analysis: D.A. Belsley (ed.) (1994) Dordrecht: Kluwer Academic Publishers. ISBN-0-7923-2356-4, pp 248, US $ 92 , 1995 .

[17]  Ian Foster,et al.  Designing and building parallel programs , 1994 .

[18]  Anthony Skjellum,et al.  Using MPI - portable parallel programming with the message-parsing interface , 1994 .

[19]  A. Atkinson.  Fast Very Robust Methods for the Detection of Multiple Outliers , 1994 .

[20]  Steve Dowers,et al.  Parallel processing for geographical applications: A layered approach , 1999, J. Geogr. Syst..

[21]  Anna Nagurney,et al.  Chapter 7 Parallel computation , 1996 .

[22]  Jack Dongarra,et al.  MPI: The Complete Reference , 1996 .