Parallel Genetic Algorithm

Parallel Genetic Algorithm（PGA）

步骤

Diversitification by a spatial population structure

The island model

• Random migration
• Random demes

The stepping stone model

• Discrete
• Migration between neighboring demes
• Demes separated in distinct population

The isolating by distance model

• Continuous distribution
• Demes are isolated by virtue of finite neighborhoodsof their members (根据有限邻域划分种群)
• For mathematical convenience it is asssumed that the position of a parent at the time it gives birth relative to that of its offspring when the latter reproduces is normally distributed（考虑到数学运算的便利，假设父代相对于子代的位置是正态分布的）

Search strategy

Local hill-climbing

• No hill-climbing
• Next ascent hill-climbing
• Steepest ascent hill-climbing

Selection and crossover

类比GA的算子，常见uniformly crossover 和 2-point crossover

Ref

Heinz Mühlenbein,Evolution in Time and Space – The Parallel Genetic Algorithm,Editor(s): GREGORY J.E. RAWLINS,
Foundations of Genetic Algorithms,Elsevier,Volume 1,1991
https://www.sciencedirect.com/science/article/pii/B9780080506845500239