B. Effects of Mirror Transformation Strategy This section explores the effects of the proposed mirror transformation strategy on the performance of the algorithm. MFEAMT and MFEA-GS are used to denote MFEA with only the mirror transformation and MFEA with only the adaptive gene selection, respectively. The MFEA-OBL is a method incorporating the classical OBL strategy defined in Eq. (2). The difference between MFEA-OBL and MFEA-MT is that MFEA-OBL directly generates the opposite point of the offspring in each generation. They are compared with MFEA and MFEA-GSMT to show the effects of the proposed mirror transformation strategy. Table II presents the experimental results of the compared algorithms. To show the sole effect of the mirror transformation, we make three comparisons, i.e., MFEA-MT vs. MFEA, TABLE I Mean and standard deviation of the approximation solution obtained by the compared algorithms on a single-objective multi-tasking test suite. MTO PROBLEM CIHS TASK GRIEWANK RASTRIGIN CIMS ACKLEY RASTRIGIN CILS ACKLEY SCHWEFEL PIHS RASTRIGIN SPHERE PIMS ACKLEY ROSENBROCK PILS ACKLEY WEIERSTRASS NIHS ROSENBROCK RASTRIGIN NIMS GRIEWANK WEIERSTRASS NILS RASTRIGIN SCHWEFEL SOEA 9.02E-01+ (5.56E-02) 4.45E+02+ (5.89E+01) 5.05E+00+ (8.70E-01) 4.39E+02+ (5.28E+01) 2.12E+01+ (3.31E-02) 4.46E+03+ (5.04E+02) 4.35E+02+ (5.79E+01) 8.75E+01+ (2.07E+01) 5.45E+00+ (9.65E-01) 2.10E+04+ (1.19E+04) 5.18E+00+ (6.76E-01) 1.19E+01+ (2.66E+00) 2.57E+04+ (9.44E+03) 4.35E+02+ (5.50E+01) 9.21E-01+ (7.29E-02) 3.91E+01+ (5.13E+00) 4.30E+02+ (5.80E+01) 4.22E+03+ (6.17E+02) MFEA 3.65E-01+ (6.73E-02) 1.81E+02+ (4.67E+01) 4.47E+00+ (7.16E-01) 2.19E+02+ (5.88E+01) 2.02E+01+ (5.28E-02) 3.69E+03+ (6.11E+02) 5.52E+02+ (9.41E+01) 8.37E+00+ (1.94E+00) 3.65E+00+ (8.93E-01) 6.80E+02+ (3.44E+02) 2.01E+01+ (1.08E-01) 1.99E+01+ (2.34E+00) 8.21E+02+ (2.29E+02) 2.55E+02+ (6.90E+01) 4.08E-01+ (9.99E-02) 2.82E+01+ (2.24E+00) 6.03E+02+ (1.18E+02) 3.68E+03+ (5.80E+02) PERFORMANCE MFEA-II 2.27E-02+ (1.24E-02) 5.57E+01+ (3.52E+01) 1.99E+00+ (3.32E-01) 8.85E+01+ (2.93E+01) 2.10E+01+ (5.14E-01) 2.12E+03+ (6.11E+02) 1.44E+02+ (3.58E+01) 4.68E-02+ (1.40E-02) 2.02E+00+ (3.49E-01) 1.57E+02+ (3.77E+01) 2.04E+00+ (4.36E-01) 2.36E+00+ (4.22E-01) 1.78E+02+ (4.76E+01) 1.19E+02+ (3.20E+01) 2.52E-02+ (1.03E-02) 1.83E+01+ (2.44E+00) 1.27E+02+ (2.76E+01) 2.04E+03 . (5.63E+02) MFEA/GHS 0.00E+00 . (0.00E+00) 0.00E+00 . (0.00E+00) 6.22E-11+ (6.12E-11) 0.00E+00 . (0.00E+00) 4.64E-01+ (1.39E+00) 1.73E+01+ (7.51E+01) 1.15E+01+ (5.13E+01) 3.32E+00+ (7.99E-01) 2.41E+00+ (3.55E-01) 6.06E+01 . (6.37E+01) 3.62E-04+ (9.75E-04) 3.37E-04+ (5.82E-04) 7.40E-01 - (2.07E+00) 1.26E-09+ (2.68E-09) 1.04E-04 - (2.22E-04) 2.40E-03+ (4.63E-03) 1.15E-10+ (3.71E-10) 1.48E-03 - (2.50E-03) MFEA-GSMT 0.00E+00 (0.00E+00) 0.00E+00 (0.00E+00) 8.88E-16 (0.00E+00) 0.00E+00 (0.00E+00) 6.91E-06 (2.14E-05) 6.36E-04 (1.54E-09) 0.00E+00 (0.00E+00) 4.14E-03 (6.84E-03) 1.54E+00 (4.19E-01) 4.62E+01 (5.96E-01) 8.88E-16 (0.00E+00) 0.00E+00 (0.00E+00) 4.61E+01 (3.71E-01) 0.00E+00 (0.00E+00) 1.69E-03 (1.94E-03) 0.00E+00 (0.00E+00) 0.00E+00 (0.00E+00) 2.40E+03 (2.24E+03) 46 IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE | NOVEMBER 2021