[BUG]: New error with termination with Time

[kwnspir]

Description of the bug

I want to run multiple optimizers for a problem and each to have a time terminator, but when i use term_dict as the examples say most of the time used for the first algorithm and almost nothing for the rest of them.
(Using multiple dicts doesnt help)

Steps To Reproduce

import time
import numpy as np
from mealpy import FloatVar, GA, PSO, DE, AOA

Objective function

def objective_func(solution):
return np.sum(solution**2)

Problem definition

problem_dict = {
"obj_func": objective_func,
"bounds": FloatVar(lb=[-100] * 30, ub=[100] * 30),
"minmax": "min",
}

term_dict = {
"max_time": 10 # seconds to run this algorithm only
}

Iter=1500
Pop=50

Run Genetic Algorithm

start_time = time.time()
optimizer_ga = GA.BaseGA(epoch=Iter, pop_size=Pop, pc=0.85, pm=0.1)
optimizer_ga.solve(problem_dict, termination=term_dict)
ga_runtime = time.time() - start_time
print("\nGenetic Algorithm:")
print(f"Best Solution: {optimizer_ga.g_best.solution}")
print(f"Best Fitness: {optimizer_ga.g_best.target.fitness}")
print(f"Runtime: {ga_runtime:.4f} seconds")

Run Particle Swarm Optimization

start_time = time.time()
optimizer_pso = PSO.OriginalPSO(epoch=Iter, pop_size=Pop)
optimizer_pso.solve(problem_dict, termination=term_dict)
pso_runtime = time.time() - start_time
print("\nParticle Swarm Optimization:")
print(f"Best Solution: {optimizer_pso.g_best.solution}")
print(f"Best Fitness: {optimizer_pso.g_best.target.fitness}")
print(f"Runtime: {pso_runtime:.4f} seconds")

Run Differential Evolution

start_time = time.time()
optimizer_de = AOA.OriginalAOA(epoch=Iter, pop_size=Pop, wf=0.8, cr=0.9)
optimizer_de.solve(problem_dict, termination=term_dict)
de_runtime = time.time() - start_time
print("\nDifferential Evolution:")
print(f"Best Solution: {optimizer_de.g_best.solution}")
print(f"Best Fitness: {optimizer_de.g_best.target.fitness}")
print(f"Runtime: {de_runtime:.4f} seconds")

Additional Information

i use mealpy 3.0.1

[thieu1995]

Hi @kwnspir,

There is a huge different between using time.time() and time.perf_counter(). We use time.perf_counter() in the search process. So when you use time.time() to wrap the outside computation, it will show the difference. Also, the time.time() will be effect when you are running multiple processes like checking website, opening word, other program.

[kwnspir]

What do you suggest me, i want each algorithm to run no more than a specified amount of time.