It will work. There's nothing particularly special about the code in custom factors; if something works elsewhere, it works in a factor.

Thanks Dan. I'll give it a go, at some point. I was under the impression that the pipeline computations need to be zippy, and that an optimizer might not be efficient enough. I guess that brings up the question, will the new Quantopian optimization API work within pipeline custom factors, and based on your response above, the answer would also be 'yes'.

Got this running, but there is a CVXPY error:

SolverError: Solver 'ECOS' failed. Try another solver.
There was a runtime error on line 44.

At this point, the custom factor idea is half-baked; I'm just trying to see how to use CVXPY within a custom factor.

import cvxpy as cvx  
from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline import Pipeline  
from quantopian.pipeline import CustomFactor  
from quantopian.pipeline.data.builtin import USEquityPricing  
from quantopian.pipeline.data import morningstar as mstar  
import numpy as np  
from quantopian.pipeline.filters import Q1500US

class OptRev(CustomFactor):  
    inputs = [USEquityPricing.close]  
    window_length = 20  
    def compute(self, today, assets, out, close):  
        x_tilde = np.mean(close,axis=0)/close[-1,:]  
        m = len(x_tilde)  
        b_t = 1.0*np.ones(m)/m  
        x = cvx.Variable(m)  
        objective = cvx.Minimize(cvx.sum_entries(cvx.square(x-b_t)))  
        constraints = [cvx.sum_entries(x) == 1, cvx.sum_entries(x_tilde*x) >= 1]  
        prob = cvx.Problem(objective, constraints)  
        prob.solve()  
        out[:] = np.squeeze(np.asarray(x.value))  
def initialize(context):  
    attach_pipeline(make_pipeline(context), 'my_pipe')  
def make_pipeline(context):  
    profitable = mstar.valuation_ratios.ev_to_ebitda.latest > 0  
    market_cap = mstar.valuation.market_cap.latest  
    top_market_cap = market_cap.top(100, mask = (Q1500US() & profitable))  
    optrev = OptRev()

    return Pipeline(columns={'optrev': optrev},screen = top_market_cap)

def before_trading_start(context,data):

    output = pipeline_output('my_pipe')  
    context.stocks = output.index.tolist()  

Got it running. Apparently, CVXPY was choking on NaNs, which are now addressed w/ np.nan_to_num (not necessarily the best approach, but easy to apply just to get the thing running).

One thing to sort out is how to apply a custom factor to a limited set of stocks. I gather that the factor may be running on all stocks in the Quantopian universe, but I think one would want to apply the optimization to a subset.