binsearch            package:GeneticsBase            R Documentation

_B_i_n_a_r_y _S_e_a_r_c_h

_D_e_s_c_r_i_p_t_i_o_n:

     Search within a specified range to locate an integer parameter
     which results in the the specified monotonic function obtaining a
     given value.

_U_s_a_g_e:

     binsearch(fun, range, ..., target = 0, lower = ceiling(min(range)),
               upper = floor(max(range)), maxiter = 100, showiter = FALSE)

_A_r_g_u_m_e_n_t_s:

     fun: Monotonic function over which the search will be performed.

   range: 2-element vector giving the range for the search.

     ...: Additional parameters to the function 'fun'.

  target: Target value for 'fun'.  Defaults to 0.

   lower: Lower limit of search range. Defaults to 'min(range)'.

   upper: Upper limit of search range. Defaults to 'max(range)'.

 maxiter: Maximum number of search iterations. Defaults to 100.

showiter: Boolean flag indicating whether the algorithm state should be
          printed at each iteration. Defaults to FALSE.

_D_e_t_a_i_l_s:

     This function implements an extension to the standard binary
     search algorithm for searching a sorted list.  The algorithm has
     been extended to cope with cases where an exact match is not
     possible, to detect whether that the function may be monotonic
     increasing or decreasing and act appropriately, and to detect when
     the target value is outside the specified range.

     The algorithm initializes two variable 'lo' and 'high' to the
     extremes values of 'range'.  It then generates a new value
     'center' halfway between 'lo' and 'hi'.  If the value of 'fun' at
     'center' exceeds 'target', it becomes the new value for 'lo',
     otherwise it becomes the new value for 'hi'.  This process is
     iterated until 'lo' and 'hi' are adjacent.  If the function at one
     or the other equals the target, this value is returned, otherwise
     'lo', 'hi', and the function value at both are returned.

     Note that when the specified target value falls between integers,
     the \em{two} closest values are returned.  If the specified target
     falls outside of the specified 'range', the closest endpoint of
     the range will be returned, and an warning message will be
     generated.  If the maximum number if iterations was reached, the
     endpoints of the current subset of the range under consideration
     will be returned.

_V_a_l_u_e:

     A list containing: 

    call: How the function was called.

 numiter: The number of iterations performed

   flag : One of the strings,  "Found", "Between Elements", "Maximum
          number of iterations reached", "Reached lower boundary", or
          "Reached upper boundary."

   where: One or two values indicating where the search terminated.

   value: Value of the function 'fun' at the values of 'where'.

_N_o_t_e:

     This function often returns two values for 'where' and 'value'. 
     Be sure to check the 'flag' parameter to see what these values
     mean.

_A_u_t_h_o_r(_s):

     Gregory R. Warnes warnes@bst.rochester.edu

_S_e_e _A_l_s_o:

     'optim', 'optimize', 'uniroot'

_E_x_a_m_p_l_e_s:

     ### Toy examples

     # search for x=10
     binsearch( function(x) x-10, range=c(0,20) )

     # search for x=10.1
     binsearch( function(x) x-10.1, range=c(0,20) )

     ### Classical toy example

     # binary search for the index of 'M' among the sorted letters
     fun <- function(X) ifelse(LETTERS[X] > 'M', 1,
                               ifelse(LETTERS[X] < 'M', -1, 0 ) )

     binsearch( fun, range=1:26 ) 
     # returns $where=13
     LETTERS[13]

     ### Substantive example, from genetics

     # Determine the necessary sample size to detect all alleles with
     # frequency 0.07 or greater with probability 0.95.
     power.fun <- function(N) 1 - gregorius(N=N, freq=0.07)$missprob

     binsearch( power.fun, range=c(0,100), target=0.95 )

     # equivalent to
     gregorius( freq=0.07, missprob=0.05)

