jso

jSO — Improved iL-SHADE with phase-based weighted mutation

jso implements the jSO algorithm: an improved variant of iL-SHADE that uses a weighted current-to-pbest mutation strategy with phase-based parameter constraints. Early phases favor exploration (lower F, higher CR); later phases shift toward exploitation. Includes LPSR.

Usage

jso(
  NP = 40,
  G = 100,
  data,
  class.name,
  structure = c("nb", "tancl", "tan", "hc"),
  edgelist = NULL,
  verbose = 25,
  ...
)

Parameters

NP · integer · default: "40"

Initial population size. Reduced linearly to 4 over evolution.

G · integer · default: "100"

Maximum generations.

data · data.frame · required

Data frame containing features and the class variable.

class.name · character · required

Name of the class variable column.

structure · character · default: "nb"

BN structure: "nb", "tan", "tancl", or "hc".

edgelist · matrix

Optional custom BN topology (edges × 2 matrix: from, to).

verbose · integer · default: "25"

Print progress every this many generations.

...

Extra arguments forwarded to bnclassify::tan_cl or bnclassify::tan_hc.

Phase-based parameter control

jSO constrains F and CR based on the current evaluation ratio (NFE / MAX_NFE):

Phase (ratio)	F constraint	CR constraint
< 0.6	F capped at 0.7	—
< 0.25	—	CR floored at 0.7
0.25–0.5	—	CR floored at 0.6

The mutation weight Fw also scales with phase: 0.7×F early, 0.8×F mid, 1.2×F late.

Return value

Object of class DE with fields: Best, BestCLL, pobFinal, CLLPobFinal, N.evals, convergence, evaluations.

Example

library(dplbnDE)
data(car)

result <- jso(
  NP = 20, G = 25,
  data = car,
  class.name = names(car)[7],
  structure = "tan",
  verbose = 5
)
print(result)