User Manual

Table of Contents

• Introduction
• Implementation and functionalities

Transition_graph_traits

Authors: F. Cazals and T.Dreyfus

Introduction

A transition graph is a graph with two types of nodes: local minima, and index one saddles (aka transitions). TG are used in two setting:

• to represent energy landscapes,

• to represent terrains. A classical case is that of nearest neighbor graphs whose vertices are endowed with a height, e.g. an estimated probability density.

Implementation and functionalities

This package provides three main classes:

• SBL::CSB::T_Transition_graph_traits_standard : a TG representation with explicit storage of the saddles as vertices of the TG. This means that each transition is linked to the connected minima via two edges.

• SBL::CSB::T_Transition_graph_traits_compressed : a TG in which the transition is associated to the edge connecting the two linked minima.

• SBL::CSB::T_Transition_graph_converter : a class to convert between two TG representations – see those mentioned in Introduction.

The serialization of such graphs is performed by functions defined in SBL::IO Transition_graph_serialize.hpp.