We all know about the djikstra's algorithm for computing single source shortest path and the complexity for the same as O(m log n) ( m being the number of edges and n the number of vertices ), this complexity results from using the heap as a data structure for the algorithm
However I recently came across a statement which says this complexity can be improved to O( m + n log n) by using another exotic type of heap as a data structure. Does anyone has any idea about this data structure which is being mentioned?.