(a) Each attraction point is associated with the tree node that is closest to it.

(b) This establishes the set of attraction points that influence each node. (Min_dist)

(c) The normalised vectors from each tree node to each source that influences the node are then found.

(d) These vectors are added and their sum is normalized again, providing the basis for locating new tree nodes.

(e) The new nodes are incorporated into the tree structure, in this case extending the main axis and initiating a lateral branch. 

(f) The neighbourhoods of the attraction points are now tested for the inclusion of tree nodes. 

(g) The neighbourhoods of the two leftmost sources have been penetrated by the new branches. Thus, the affected attraction points are removed. 

(h) The tree nodes closest to these points are now identified. Beginning the next iteration of the algorithm.