Backtracking

Araba Aman
[email protected]

Haramaya University
 Backtracking?
▪Many problems with searching a set of solutions or which ask for
an optimal solution satisfying some constraints can be solved using
the backtracking formulation.
▪In many applications of the backtrack method, the desired solution
is expressible as an n-tuple (xi,…,xn),where the xi are chosen from
some finite set Si.
▪Often the problem to be solved calls for finding one vector that
maximizes (or minimizes or satisfies) a criterion function P(x1, x2,
.., xn). Some time its seeks all vectors that satisfy P.
▪Many of the problems we solve using backtracking require that all
the solutions satisfy a complex set of constraints.

1/16/2022 qubee-tp.com 2
Cont.
▪For any problem these constraints can be classified into two
categories: explicit and implicit.
▪The explicit constraints depend on the particular instance I of
the problem being solved.
▪All tuples that satisfy the explicit constraints define a possible
solution space for I.
▪The implicit constraints are the rules that determine which of the
tuples in the solution space of I satisfy the criterion function.

▪Thus implicit constraints describe the way in which xi must

1/16/2022 qubee-tp.com 3
Cont.
▪Explicit constraints are rules that restrict each Xi to take on values
only from a given set.

1/16/2022 qubee-tp.com 4
E.g.,
▪8-queens problem: the problem is place 8 queens on 8X8 chess
board, such that there is no attack. The attack happens when two
queens placed on the same row, column and diagonal.
Constraints
▪ Explicit constraints
▪ – Si={1,2,…,8}
▪ – So, solution space consists of 88 8-tuples
▪ Implicit constraints
▪ – No two xi‘s can be the same (By this, solution space reduced
from 88 to 8!)
▪ – No two queens can be on the same diagonal

1/16/2022 qubee-tp.com 5
Cont.

1/16/2022 qubee-tp.com 6
Cont.
▪ Greedy Algorithm:

▪ Dynamic Programming:

▪ Backtracking:

1/16/2022 qubee-tp.com 7
▪ In backtracking, the basic idea is to build up the solution vector one
component at a time
▪ and to use modified criterion functions Pi (x1,…xi) (sometimes called
bounding functions) to test whether the vector being formed has any
chance of success.
▪ The major advantage of this method is this: if it is realized that the partial
vector (x1, x2,….., xi) can in no way lead to an optimal solution, then
mi+1……mn possible test vectors can be ignored entirely.

▪ Backtracking algorithms determine problem solutions by systematically

▪ This search is facilitated by using a tree organization for the solution space.
1/16/2022 qubee-tp.com 8
Cont.
▪ N-QUEENS problem; it is generalization of 8-queens problem. That is
placing n queen on nXn chess board without attacks.
▪ The solution space consists of all n! permutation of n-tuples. The tree is
called permutation tree.
▪ Edges are labeled by possible values of Xi.
▪ Edges from level 1 to level 2 nodes specify the values for x1.
▪ Edges from level i to i+1 are labeled with values of xi.
▪The solution space is defined by all the paths from the root node
to a leaf node.
▪ For e.g., If n=4, then there will be 4! =24 leaf nodes in the tree.

1/16/2022 qubee-tp.com 9
1/16/2022 qubee-tp.com 10
1/16/2022 qubee-tp.com 11
1/16/2022 qubee-tp.com 12
▪ Let G be a graph and m be a given positive integer. We want to discover
whether the nodes of G can be colored in such a way that no two adjacent
nodes have the same color yet only m colors are used - this is called m-
colorability decision problem.
▪ The m-colorability optimization problem asks for the smallest integer m for
which the graph can be colored. This integer is called chromatic number of
the graph.
▪ Chromatic number of this graph is 3.

1/16/2022 qubee-tp.com 14
Cont.
▪ E.g., if we apply graph coloring to map coloring, region of the map becomes
node, and if two regions are adjacent, then corresponding nodes are joined by
edges.
▪ E.g., Map regions, and corresponding graph nodes and edges.

▪ For many years it was known that five colors are enough to color any map, but no
map that required more than four colors have ever found.

1/16/2022 qubee-tp.com 15
Cont.
▪Suppose we represent a graph by its adjacency matrix G[1:n,
1:n], where G[i, j]=1 if (i, j) is an edge of G, and G[i, j]=0
otherwise.
▪The colors are represented by the integers 1, 2, 3, .. , m and the
solutions are given by the n-tuple (x1, x2, …, xn), where xi is
the color of node i.
▪The underlying state space tree used is a tree of degree m
and height n+1.
▪That means, each node at level i has m children corresponding to
the m possible assignments to xi, 1<=i<=n.
▪ Nodes at level n+1 are leaf nodes.

1/16/2022 qubee-tp.com 16
Cont.

▪NB: if I choose x1=1 for first vertex out of three colors (1, 2, 3),
then still I can choose x2 of vertex two any one of the three
colors.
1/16/2022 qubee-tp.com 17
0 2 0 3
2 0 4 1
9 8 0 3
0 0 2 1

1/16/2022 qubee-tp.com 18
Cont.
▪Means that, at each stage, or vertex I have three options (colors)
to choose from.
▪ The total time: O(nmn)
So, how to apply DP to Graph Coloring?
▪ Apply bound condition (constraint) if can be colored, choose xi
for a vertex i. Stop expanding the tree otherwise.
▪Try coloring vertex next to vertex i with other color, keep on
expanding the tree until you find right answer for xn. Where
1<=i<=n.
▪ The constraint is: not to adjacent vertices have the same color.

1/16/2022 qubee-tp.com 19
E.g.,
▪ Let take n=3 and m=3 (R, G, B) or say Red, Green and Blue colors.
Steps:
▪ Draw a root node.
▪ Know that, X1 holds color value of vertex 1.
At start X1 three options; Red, Green, Blue.
Choose X1 = Red.
▪ Then expand the tree to third level, for second vertex
Find X2 out of the three. Try X2=Red, since vertex 1 is adjacent to vertex 2 X2
can't take Red color, then stop expanding that way. Then choose Green, check
if colorable, yes. It can be colored.
1. a

▪ Then move this direction, expand tree to fourth level. Find X3, try the two
colors Red, Green one at a time. If colorable do it, else try third color Blue.
1/16/2022 qubee-tp.com 20
 Homework
▪ Find at least the first four solutions

1/16/2022 qubee-tp.com 21