Enumeration solvers

OptimalApplication.optimalportfolio_enumerateFunction
optimalportfolio_enumerate(mkt::SameCostsMarket) -> X, v

Produce the optimal portfolio for SameCostsMarket defined by mkt. Very slow; you probably should use applicationorder_list or applicationorder_heap instead.

julia> mkt = SameCostsMarket([0.2, 0.5, 0.1, 0.6, 0.1], [1, 4, 9, 1, 8], 2);

julia> optimalportfolio_enumerate(mkt)
([2, 3], 2.7)
source
optimalportfolio_enumerate(mkt::VariedCostsMarket) -> X, v

Produce the optimal portfolio for VariedCostsMarket defined by mkt. Very slow; you probably should use optimalportfolio_dynamicprogram or optimalportfolio_branchbound or optimalportfolio_simulatedannealing instead.

julia> mkt = VariedCostsMarket([0.2, 0.5, 0.1, 0.6, 0.1], [1, 4, 9, 1, 8], [2, 4, 2, 5, 1], 8);

julia> optimalportfolio_enumerate(mkt)
([2, 5, 3], 3.24)
source
OptimalApplication.optimalportfolio_branchboundFunction
optimalportfolio_branchbound(mkt::VariedCostsMarket; maxit=10000, verbose=false) -> X, v

Use the branch-and-bound algorithm to produce the optimal portfolio for the VariedCostsMarket defined by mkt. Intractable for large markets.

julia> mkt = VariedCostsMarket([0.2, 0.5, 0.1, 0.6, 0.1], [1, 4, 9, 1, 8], [2, 4, 2, 5, 1], 8);

julia> optimalportfolio_branchbound(mkt)
([2, 3, 5], 3.24)
source
OptimalApplication.optimalportfolio_dynamicprogram_slowFunction
optimalportfolio_dynamicprogram_slow(mkt, δ) -> X, v

Uses a dynamic program to produce the optimal portfolio X with valuation v, for the VariedCostsMarket defined by mkt. The valuation of X differs from that of the global optimum by no more than δ.

julia> mkt = VariedCostsMarket([0.2, 0.5, 0.1, 0.6, 0.1], [1, 4, 9, 1, 8], [2, 4, 2, 5, 1], 8);

julia> optimalportfolio_dynamicprogram_slow(mkt, 1e-6)
([3, 5, 2], 3.24)

This dynamic program iterates on portfolio valuations and is an "exact" form of the approximation algorithm implemented in this package as optimalportfolio_fptas. For the vast majority of problem instances, optimalportfolio_dynamicprogram is a faster way to obtain an exact solution, but this one technically can be used to obtain the global optimum by setting δ = eps().

source