Jianjun Hu, Erik D. Goodman, Kisung Seo

1
HFC a Continuing EA Framework for Scalable
Evolutionary Synthesis

• Jianjun Hu, Erik D. Goodman, Kisung Seo
• Zhun Fan, Ronald C. Rosenberg
• Genetic Algorithm Research Applications Group
  (GARAGe)
• Michigan State University

2
Evolutionary Synthesis the problem
Input gtBuilding blocksgtFunction
Specification(Evaluation function)gtEA settings
(operators, parameters)
Output Design solutions
Evolution

• Topology innovation How many and what types of
  building blocks, How to connect them?
• Parameter innovation How to size the elements?

3
Sustainable scalable evolutionary synthesis
Definition and Reality

• Definition the capability to obtain better
  results or solve larger scale problems when given
  more computing resources
• Reality not sustainable and not scalable
• Bloating and aging problem (e.g.Innovation in 50
  generations in GP)
• Demand for huge population size (GP)
• Premature convergence and local optima (all EA)
• But biological evolution is sustainable and
  scalable!

4
Sustainable scalable evolutionary synthesis
two types of obstacles

• Two types of obstacles
• 1. Convergent nature of current EA framework
• one of major factors leading to
• GP aging phenomenon
• premature convergence
• dependence on huge population size
• 2. Non-scalable compositional mechanisms for
  topological and parametric evolution
• possible solutions modularity, hierarchical
  organization, biological developmental
  principles
• This paper
• addresses the EA convergence problem

5
Comparison with biological evolution how to
achieve sustainable evolution

• Biological evolution
• Almost infinite population size
• Simultaneous evolution of all levels of
  organisms bacteria coexist with humans
• Fair competition different levels of organism
  coexist in different niches
• Sustainable innovation possible

• Artificial evolution (EA)
• Limited population size
• Focusing on current high-fitness individuals
• Unfair competition highly-evolved individuals
  compete with new offspring of low fitness
• Innovation capability rapidly depleted

6
Sustainable evolution example comparison

• Unsustainable EA how generations of solutions
  are evolved?

• Sustainable education system how generations of
  talents are educated?

7
Assembly-line structured continuing EAs
8
HFC-EA framework
9
System synthesis problem eigenvalue placement
-2.0 -3.3 -2.0 3.3 -7.5 -4.5 -7.5 4.5 -3.5 -12.0 -3.5 12.0 -3.4 -12.0 -3.4 12.0 -10.0 -8.0 -10.0 8.0 -2.063 -3.005 -2.063 3.005 -7.205 -4.581 -7.205 4.581 -3.580 -11.835 -3.580 11.835 -3.334 -12.526 -3.334 12.526 -9.997 -8.040 -9.997 8.040
Max distance error 0.530 Average distance error 0.272
2 hours 2 hours
10
Experiment result sustainability robustness
Dynamic system synthesis problem with
simultaneous topology and parameter search
11
Experiment result handling GP aging problem
10 eigenvalue dynamic system synthesis problem
12
Experiment result small population size works
equally well
10-parity problem with function set and, xor,
or, not)
13
Why HFC works the explanation
14
Conclusion Hierarchical Fair Competition
Principle for EA

• HFC is very effective in evolutionary synthesis
• Simultaneous evolution at all (fitness) levels,
  from the random population to best individuals
• Avoid premature convergence by allowing emergence
  of new optima rather than trying to jump out of
  local optima
• Allow use of strong selection pressure without
  risk of premature convergence
• Small population size also works

15
Ongoing future work

• Developing single population HFC (CHFC) (with
  continuous level segmentations) algorithm to
  achieve sustainable evolution
• Developed HFC-enhanced multi-objective EAs
  (GECCO2003)
• To develop hybrid parallel-HFC GA/GP system where
  each deme is implemented as a CHFC population
• To develop multi-level system synthesis from
  framework evolution to complete solutions

16
Generalization of HFC-EA Framework

• A generic framework for continuing sustainable
  evolutionary computation (GA, GP, ES, )
• Especially good for Evolutionary Synthesis for
  Sustainable topological innovation which has
  extremely rugged fitness landscape.
• Especially effective for problems with
• High multi-modality
• Strong tendency of premature convergence
• Requirement on robustness
• Requirement on adaptation in dynamic environment
• Also applicable for artificial life evolution

17
Scaling Mechanism of HFC

• A natural parallel evolutionary computation
  model. Better than island parallel model
• Hybridizing with single-population HFC EAs (Each
  deme is a sustainable HFC subpop)
• Natural hybridizing with explicit hierarchical
  building block discovery mechanisms
• Allow using small population size and longer time
  to achieve good results
• No restart to waste computing effort