Blood Inheritance Optimization: An Evolutionary Search Method

Summary

Blood Inheritance Optimization (BIO) is a population-based method for searching parameter spaces. It assigns each candidate one of four blood-type labels and uses an inheritance table to choose a label for each new candidate based on its two parents. Parents are selected with a quadratic probability distribution that favors higher-fitness solutions. The child’s label determines how each parameter is formed: it may copy the best-known value, mutate a parent value toward a range edge, move it toward the current best, or reflect it across the parameter range.

The method maintains a global best and carries strong candidates into later generations. The article reports tests on benchmark functions, describing good convergence on medium and high-dimensional problems, but a tendency to get stuck at local optima on low-dimensional ones. It provides no detailed performance figures in the supplied text, and its results concern general optimization benchmarks rather than trading strategies. The author also cautions that implementations of canonical algorithms may be modified, which limits direct comparisons.

Key ideas

• BIO assigns each candidate a label that determines how its parameters are inherited and mutated.
• Parent selection favors higher-fitness candidates through a quadratic probability distribution.
• The four mutation strategies range from copying the best-known values to reflecting parent values across parameter bounds.
• The article reports stronger convergence on medium and high-dimensional test functions than on low-dimensional problems, where local optima can trap the search.
• The reported tests use benchmark functions and do not establish performance in financial markets.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.