Performance and Optimization

Sometimes optimization process takes a lot of time to generate single lineup. It usually happens in mlb and nfl because all teams plays in same day and each team has a lot of players and total number of players used in optimization is >500. In this case a good approach is to remove from optimization players with small fppg value and big salary.

optimizer = get_optimizer(Site.DRAFTKINGS, Sport.BASEBALL)
for player in optimizer.players:
    if player.efficiency == 0:
for lineup in optimizer.optimize(10):

Optimizer parameters tuning

For some special cases with a lot of different constraints you can try to tune solver parameters. pydfs-lineup-optimizer uses PuLP library for solving optimization problem, by default it uses CBC solver so you can try to change default parameters. You can find list of available parameters here. This is example of tuning parameters:

from pulp.solvers import PULP_CBC_CMD
from pydfs_lineup_optimizer import get_optimizer, Site, Sport
from pydfs_lineup_optimizer.solvers.pulp_solver import PuLPSolver

class CustomPuLPSolver(PuLPSolver):
    LP_SOLVER = PULP_CBC_CMD(threads=8, options=['preprocess off'])

optimizer = get_optimizer(Site.DRAFTKINGS, Sport.BASEBALL, solver=CustomPuLPSolver)

You can try to change solver as well for any solver that PuLP support: glpk, cplex, gurobi etc.