the outcome of a B&B More...

#include <PBeFS/bnb/BranchAndBound.h>

Public Member Functions
bool	feasible (void) const noexcept
	true if a feasible value is known

Public Attributes
BBStatus	status = BBStatus::infeasible
	how the search ended
std::optional< Node >	node
	the incumbent node
V	value = BBValueTraits< V >::none()
	its value
V	bound = BBValueTraits< V >::none()
	the bound
V	gap = BBValueTraits< V >::infinity()
	the gap
BBStats	stats
	the counters
std::vector< BBLeaf< Node, V > >	frontier
	the leaves, if kept

Detailed Description

template<class Node, class V = double>
struct PBeFS_di_unipi_it::BBSolution< Node, V >

the outcome of a B&B

value is the incumbent value and bound a bound on the optimum, both in the sense of the problem, and the gap, value minus bound in absolute value, is valid whatever the status; value is none() of BBValueTraits if no feasible value is known, and node is the incumbent node, empty if no feasible node was found (even when an initial incumbent value was given).

With BBOptions::keep_frontier, frontier holds the leaves of the tree explored, in no particular order: the nodes pruned, feasible or infeasible, and those left open by a limit. Their subtrees cover the subtrees of the nodes the search started from, so that a search started from them [see BranchAndBoundSearch::solve()] explores the same region, e.g., after a change of the data that leaves the branching valid.