Multiobjective Memetic Estimation of Distribution Algorithm Based on an Incremental Tournament Local Searcher

Algorithm 2

The details of MMEDA.

Step  1. Initialization: Set 𝑡=0, generate initial population P(0) and evaluate it.

Step  2. Update non-dominated population: Identify non-dominated solutions from population P(t), and update archive N(t)

by 𝜀 dominance.

Step  3. Termination: If 𝑡>𝐺_max is satisfied, export N(t) as the output of the algorithm, and Stop, else go to step 3.

Step  3. ITLS: Perform incremental tournament local searcher to generate 𝑁_𝑠 new solutions 𝐍_𝐬(t).

Step  4. Modeling: Perform the (m-1)-d local PCA to partition P(t) into 𝐾 disjoint clusters 𝑆¹,𝑆²,…,𝑆^𝐾. For each cluster 𝑆^𝑗,

build model (2) by (4) and (5).

Step  5. Sampling: Sample new population O(t) from model (2) and evaluate O(t).

Step  6. Update current population: select 𝑁 solutions from 𝐏(𝐭)∪𝐎(𝐭)∪𝐍_𝐬(𝐭) to create 𝐏(𝐭+𝟏).

Step  7. set 𝑡∶=𝑡+1 and go to Step 2.