A Decomposition-Based Harmony Search Algorithm for Multimodal Multiobjective Optimization
Algorithm 1
MOEA/D-HSA.
Input: population size NP, decision variable dimension n, algorithm parameters: HMS, HMCR, PAR, bandwidth (BW), pr, t, the maximum number of evaluation functions MaxFES,
Output: the set TP.
(1)
Initialize population P and weight vector W;
(2)
Calculate the fitness Fit and ideal point ;
(3)
Set FES = NP;
(4)
for i=1:NP do
(5)
Set = [];
(6)
end for
(7)
for i = 1:NP do
(8)
, where ;
(9)
end for
(10)
while FES < MaxFES do
(11)
Set ;
(12)
for i = 1:|T| do
(13)
Generate offspring by Algorithm 2 and elite learning strategy;
(14)
Update ideal point and subpopulation by Algorithm 3, where ;