\title{A Simulated Annealing Algorithm for Maximum 
       Likelihood Pedigree Reconstruction}

\author{Anthony Almudevar} 



{\bf Abstract} 
The calculation of maximum likelihood 
pedigrees for related organisms using genotypic data is considered.
The problem is  formulated so that the domain of optimization 
is a permutation space. This is a feature shared by the travelling 
salesman problem, for which simulated annealing 
is known to be effective. Using this technique it is found that 
pedigrees can be reconstructed with minimal error using
genotypic data of a quality currently realizable. This can 
be done without any {\it a priori} age or sex information. 
For smaller numbers of individuals a method of efficiently 
enumerating all admissible pedigrees of nonzero likelihood is given.