The population iteration of LBMBC-MOEA algorithm begins with the initial population, so we need to determine the initial population first. In general problems, the chromosomes in the initial population can be generated randomly. However, in LBMBC problem, the nodes in the solution corresponding to a chromosome must form a CDS. Therefore, we design a population initialization algorithm to generate the first generation that meets the requirement. In the population initialization algorithm, we first use an existing CDS based backbone construction algorithm [
26] to create the first chromosome
. Assuming that the size of the population is
, we will generate other
chromosomes in the population according to the first chromosome
. A new chromosome is generated by adding one node to the node set corresponding to
each time. Since the node set corresponding to
is a CDS, it is guaranteed that the node set generated by this method is also a CDS. When selecting a new node to the node set corresponding to
, according to the optimization objectives of LBMBC problem, we select one of the optimization objectives, for example the multicast ability of the node, to sort the other nodes in the network. If the optimization objective is to minimize, the nodes are arranged in ascending order, and if it is to maximize, the nodes are arranged in descending order. We add one node by the order into the corresponding node set of
each time to generate a new non-duplicated chromosome. If the number of chromosomes is not enough after the above steps, repeat the above steps from chromosome
until there are
non-duplicated chromosomes in the initial population.