A study was made of the problem of locating M facilities
on a connected grid graph, so that M is the minimum and so
that every demand node on the graph is within given distance
K of one of these M facilities. We call this problem briefly
the G(N,K,M) problem, with N denoting the total number
of demand nodes. An algorithm for solving this problem by using backtrack
technique is presented in this thesis. A heuristic algorithm
is also present; although the resulting M is not always minimum,
it tends to be near minimum. The advantage over the
backtrack algorithm is that the heuristic algorithm operates
very quickly. Algorithms represented in this thesis are programmed in
the Pascal language for the Univac 1100 computer at Florida
Atlantic University, Boca Raton, Florida.