Skip to content

pymoo/unread-history

Rule unread-history
Pack pymoo
Severity info
Fix unsafe, with xpd check --fix --unsafe-fixes
Library versions every version

save_history=True keeps a deep copy of the algorithm for every generation, which made a run 2.6 times slower and its result 150 times larger, and this file never reads it. Drop save_history=True.

res = minimize(problem, algorithm, ("n_gen", 200), seed=1, save_history=True)
res = minimize(problem, algorithm, ("n_gen", 200), seed=1)
print(res.F)

The rule reports save_history=True given to minimize outside every function, in a file that never names history and does not use pickle, dill or joblib. The option keeps a deep copy of the algorithm for every generation. When another file reads the history of the run, the report is wrong.

NSGA2, 100 solutions, 200 generations Time Pickled result F
save_history=True 1.17 seconds 20.3 MB the same
without it 0.45 seconds 0.13 MB the same

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

The unsafe fix drops save_history=True, so res.history is empty.

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

save_history=True keeps a deep copy of the algorithm for every generation, and this file never reads history. A run of 200 generations with 100 solutions took 2.6 times as long, and its result pickled to 20.3 MB against 0.13 MB, with the same F. Drop save_history, or record the values you need in a Callback.

Browse the other rules of the pymoo pack.