Implementation of the Branch-and-Bound method for solving Integer Linear Programming problems.
A serial executable is obtained by using your favourite C++ compiler to compile and link the downloaded files. For example, using g++:
g++ bnb.cpp -o bnb.exe
After done compiling, execute the 'bnb' program passing the ILP problem data file as an argument, like this:
./bnb.exe entrada/DadosXX.esd
If no argument is passed, the program will read the ILP problem from the user itself. Both ways will use the default search and print every node.
The default search is DFS(Depth-First Search) with backtracking, giving priority to '<=' constraints. But there are a few arguments that you can choose before you start the search.
./bnb.exe entrada/DadosXX.esd ARG1 ARG2 ARG3
ARG1: Search type, (0) for DFS with backtracking, and (1) for BFS(Breadth-First Search).
ARG2: Son priority, (0) for '<=', and (1) for '>=' constraints.
ARG3: Specifies how often the program will print the nodes on the screen. '1000' will print every 1000 new nodes, including the root node.
If ARG2 or ARG3 are not specified, the program will set priority to '<=' constraints and print every new node, respectively.
To pass the problem as an argument to the calculator, it needs to be in the specific format below, so that the Data class can read the file correctly:
NAME: name of your file or problem.
p_t: especifies the problem type -> (0) for Maximization, (1) for Minimization.
n_v: number of variables in the Objective Function (OF).
n_c: number of constraints.Z = XX XX XX.. (OF)
XX XX XX.. < RHS (Constraint 1)
XX XX XX.. = RHS (Constraint 2)
XX XX XX.. > RHS (Constraint 3)
'<' is read as '<='.
'>' is read as '>='.
RHS holds the value of the Right-Hand Side of that constraint.
Each XX specifies the coefficients of the variables from the OF.