Research Article

Uniformity-Comprehensive Multiobjective Optimization Evolutionary Algorithm Based on Machine Learning

Table 6

The IGD comparison results of MOEA-UCML with two initialization improved multiobjective optimization algorithms on DTLZ and WFG problem sets.

ProblemCMOPSOMOEA/D-AAWNMOEA-UCML

DTLZ41.4430e − 1 (2.8e − 1)−4.8771e − 1 (3.5e − 1)−1.0003e1 (2.0e1)
DTLZ54.5173e − 3 (3.4e − 4)−3.2439e3 (1.2e4)+5.8972e − 3 (2.4e − 4)
DTLZ64.3734e3 (6.1e5)+1.5703e − 1 (3.0e − 1)−5.8490e − 3 (3.0e − 4)
WFG43.8183e − 1(3.9e − 3)−2.1752e1 (1.4e2)+2.7556e − 1 (9.7e − 3)
WFG57.7588e − 1 (1.8e − 3)−2.2549e1 (1.2e2)+2.7339e − 1 (9.4e − 3)
WFG68.6534e − 1 (1.2e − 2)−2.8890e − 1 (2.1e − 2)−2.6029e1 (1.9e2)
WFG72.9628e − 1 (2.2e − 3)−2.0988e − 1 (5.4e − 2)2.0147e1 (5.0e3)
WFG84.0653e − 1 (8.1e − 3)−3.1907e − 1 (4.9e − 2)3.1162e1 (8.0e3)
WFG97.3108e − 1 (3.9e − 2)−2.6916e − 1 (5.1e − 2)−2.0381e1 (7.8e3)
8/1/04/3/2

The bold values indicate the optimal performance results of all algorithms for each test function.