The problem considered here is an unconstrained
minimization problem, which has the general form:
f x( ) min → , n x ∈ R , (P)
where : n f R R → is a bounded below continuous strictly
unimodal function.
We use the following definition of strict unimodality.
Definition. Let D be a bounded closed convex set in n R .
Function f : D R → is strictly unimodal over set D iff
for any segment Δ ⊂ D # min 1 Arg f x x { } ( ) ∈ Δ = ,
where « # A» is the cardinality of set A.
The multidimensional bisection method (Baushev and
Morozova, 2007) allows to solve constrained
minimization problem when the feasible region is ndimensional
simplex. This method generalizes a onedimensional
bisection method for the case n>1 using a
recursive procedure. This paper will present an extension
of the multidimensional bisection method for solving
problem (P). This method does not require a
differentiability of function f, and is guaranteed to
converge to the minimizer for the class of strictly
unimodal functions.