Let's say I have formulated an integer linear programming (ILP) problem with the objective function $$F(X)=V(T,X)-C(t,X),$$ where $V(T,X)$ is the payoff of portfolio, and $C(t,X)$ is the initial cost of portfolio, $0<t<T$ is the calendar time, then I have setup a system of constraints and found the optimum solution $X=(x_1, x_2, \ldots, x_n)$, where $x_i$ is the number of units of an $i$-th asset in the portfolio, with $x_i>0$ for buying, $x_i<0$ for short selling.
Now I'd like to extend the system of constraints and add new constraint on the initial cost $C(t,X)$. Let's say $C(t,X)\le c$, where $c$ can be either a positive number or zero, or even negative number. I think that theoretically I can find the optimum solution $X$ with the constraint $C(t,X)\le c$.
In the study (Bartoňová M., 2012) was demonstrated the usage of zero-cost option’s strategy in hedging of sales. But on the page 125 the author conclude:
And are they really zero-cost? As for initial fee, than yes. It is necessary to take into consideration that there is necessary general agreement with bank for option trading. It must be covered by collateral. There are also costs of contract processing, expert's opinions for assets evaluation, opportunity costs influencing of pledge, also of call option sale... Any zero-cost options are not really zero.
My question: Can I assume than an ivestor can use the money received from the sale of some contracts to buy of other contracts in the portfolio? Can I realize the optimal porfolio with the zero or negative initial cost on a market?
Peter Carr and Dilip Madan. Towards a theory of volatility trading. In R. Jarrow, editor, Volatility , pages 417-427. Risk Publications, 1998.