Algorithm 2: Level-based learning and knowledge transfer

Input: the source task particle swarm Ps, the target task particle swarm Pt, number of levels L, random mating probability rmp, level size LS, control parameter ϕ.   /* Initialization */ OP = {} For i = {L, …, 2} do For j = {1, …, LS} do xnew = xi,j; vnew = vi,j; If i ! = 1 then Select two level from top (i−1) levels: l1, l2; Else l1 = l2 = 1; End If If l1 > l2 then Swap (l1, l2); End If If rand < rmp then $\mathrm{Select}\mathrm{two}\mathrm{particles}\mathrm{from}{l}_1,l_2\mathrm{respectively}\mathrm{in}{P}_{\mathrm{t}}:x_{l_1}^{\prime },x_{l_2}^{\prime };$ Else $\mathrm{Select}\mathrm{two}\mathrm{particles}\mathrm{from}{l}_1,l_2\mathrm{respectively}\mathrm{in}{P}_{\mathrm{s}}:x_{l_1},x_{l_2};$ End If Update vnew and xnew according to Equations (3) and (4); particlenew = {vnew, xnew,}; OP = OP ∪ particlenew;

Output: OP