4.2.3. New evolutionary algorithm for the MCPP
Because ALA-EA uses all possible edges for an encoding
at network problems, for applying it to MCPP we
should be able to encode the problem to satisfy this condition.
Fig. 1 shows the encoding method of ALA-EA for
MCPP. In there, all items are represented as they can be
assigned to each container and all gene values are set to
‘0’ at the initialization process. Therefore, the edge numbers
correspond to the decision variable xi,j.
For the decoding, ALA-EA first selects the edge e with
the lowest gene value and checks whether the edge e can
be included in the assignment plan. If possible, then the
decision variable xi,j of the edge e is set to ‘1’, otherwise
‘0’. And then the edge considered is eliminated from set
A, which indicates the set of all possible edges. This process
is repeated until set A empties. Finally, a feasible solution is
generated through this process.