Skip to main content

Table 2 Settings for the standard algorithms in detail

From: Modeling metabolic networks in C. glutamicum: a comparison of rate laws in combination with various parameter optimization strategies

Algorithm

Population

Mutation

Crossover

Selection

Monte Carlo

50

no

no

Best

Hill Climber

1, 10, 25, 50, 100, 250

Fixed Step σ = 0.2, p m = 1

no

Best

binGA

250

one-point, p m = 0.1

one-point, p c = 0.7

Tournament, group of 8

realGA

250

global, p m = 0.1

UNDX, p c = 0.8

Tournament, group of 8

stdES

(5, 25)

global, p m = 0.8

discrete one-point, p c = 0.2

Best

cmaES

(5,+25)

CMA, p m = 1

no

Best

SA

250

linear annealing schedule, α = 0.1, initial T = 5

Best

DE

100

current-to-best/1, λ = F = 0.8, CR = 0.5

PSO

100

star topology, ϕ1 = ϕ2 = 2.05, χ = 0.73

  1. This overview lists the standard settings for the optimization procedures used to infer the parameters in the differential equation systems. The algorithms DE and PSO do not follow the general scheme of mutation and crossover. The Tribes algorithm is not listed here as it was designed to be a settings-free derivative of PSO. The cmaES is used with and without elitism (plus or comma strategy). For details, see Methods and [see Additional file 1].