Skip to content

pymoo/decision-weights

Rule decision-weights
Pack pymoo
Severity warning
Fix unsafe, with xpd check --fix --unsafe-fixes
Library versions every version

The decision method reads these weights another way than the code means, so the larger weight counts less. Pass 1 / weights, as the pymoo guide does, when a larger weight should count more.

i = ASF().do(F, np.array([0.2, 0.8])).argmin()
j = PseudoWeights(np.array([2, 8])).do(F)
i = ASF().do(F, 1 / np.array([0.2, 0.8])).argmin()
j = PseudoWeights(np.array([0.2, 0.8])).do(F)

The rule reports weights that the decision method reads in another way than the code means.

Code How pymoo reads the weights What it picked
ASF().do(F, np.array([0.2, 0.8])), with weights written out and not all the same ASF divides each objective by its weight, so the larger weight counts less. The pymoo guide passes 1 / weights. the solution (0.1, 0.9), where 1 / weights picked (0.9, 0.1)
a weight above 1 written out in the argument of PseudoWeights, such as (2, 8) PseudoWeights compares the weights with shares that sum to 1. row 0 of five solutions, where (0.2, 0.8) picked row 3

xpd check --fix --unsafe-fixes applies the fixes of this rule. Check the result, because an unsafe fix can change what the code does.

unsafe for ASF: 1 / weights, which picks another solution, the one the code means when a larger weight stands for a more important objective. None for PseudoWeights.

xpd prints this message with the issue. A name that starts with $ stands for code from your file.

ASF divides each objective by its weight, so the larger weight counts less. With the weights (0.2, 0.8), ASF picked the solution (0.1, 0.9), and with 1 / weights it picked (0.9, 0.1). Pass 1 / $W, as the pymoo guide does, if a larger weight means a more important objective.

Browse the other rules of the pymoo pack.