The function updates an existing cluster. This is done as follows. First the
number of patterns in the cluster is retrieved from the cluster table. Then
the pattern representing the cluster is updated according to the formula:
((n/m*cluster_value) + (1/m*pattern), where n/m = n/(n+1) with n is the
number of patterns in the cluster and pattern the pattern to be added to the
cluster. Next, the cluster table is updated by increasing the pattern counter
and adding the pattern number to the table.
NOTE
In order for this function to work properly, the variables have to be
initialized using cluster_pao_init.