-
Notifications
You must be signed in to change notification settings - Fork 43
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
5 changed files
with
254 additions
and
6 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,173 @@ | ||
| # This file implements a toy bin packing model for Node Finalizer. It solves an instance with | ||
| # three items where any two of them fits into a bin but the three together do not. Pricing is | ||
| # solved by inspection onn the set of six possible solutions (three singletons and three pairs) | ||
| # which gives a fractional solution at the root node. Then node finalizer function | ||
| # "enumerative_finalizer" is called to find the optimal solution still at the root node and | ||
| # avoid branching (which would fail because maxnumnodes is set to 1). | ||
| # If "heuristic_finalizer" is true, then it allows branching and assumes that the solution found | ||
| # is not necessarily optimal. | ||
| CL.@with_kw struct EnumerativeFinalizer <: ClA.AbstractOptimizationAlgorithm | ||
| optimizer::Function | ||
| end | ||
|
|
||
| function ClA.run!( | ||
| algo::EnumerativeFinalizer, env::CL.Env, reform::ClMP.Reformulation, input::ClA.OptimizationInput | ||
| )::ClA.OptimizationOutput | ||
| masterform = ClMP.getmaster(reform) | ||
| _, spform = first(ClMP.get_dw_pricing_sps(reform)) | ||
| cbdata = ClMP.PricingCallbackData(spform, 1) | ||
| isopt, primal_sol = algo.optimizer(masterform, cbdata) | ||
| result = ClA.OptimizationState( | ||
| masterform, | ||
| ip_primal_bound = ClA.get_ip_primal_bound(ClA.getoptstate(input)), | ||
| termination_status = isopt ? CL.OPTIMAL : CL.OTHER_LIMIT | ||
| ) | ||
| if primal_sol !== nothing | ||
| ClA.add_ip_primal_sol!(result, primal_sol) | ||
| end | ||
| return ClA.OptimizationOutput(result) | ||
| end | ||
|
|
||
| function node_finalizer_tests(heuristic_finalizer) | ||
|
|
||
| function build_toy_model(optimizer) | ||
| toy = BlockModel(optimizer, direct_model = true) | ||
| I = [1, 2, 3] | ||
| @axis(B, [1]) | ||
| @variable(toy, y[b in B] >= 0, Int) | ||
| @variable(toy, x[b in B, i in I], Bin) | ||
| @constraint(toy, sp[i in I], sum(x[b,i] for b in B) == 1) | ||
| @objective(toy, Min, sum(y[b] for b in B)) | ||
| @dantzig_wolfe_decomposition(toy, dec, B) | ||
|
|
||
| return toy, x, y, dec, B | ||
| end | ||
|
|
||
| @testset "Optimization algorithms that may conquer a node" begin | ||
|
|
||
| call_enumerative_finalizer(masterform, cbdata) = enumerative_finalizer(masterform, cbdata) | ||
|
|
||
| coluna = JuMP.optimizer_with_attributes( | ||
| CL.Optimizer, | ||
| "default_optimizer" => GLPK.Optimizer, | ||
| "params" => CL.Params( | ||
| solver = ClA.TreeSearchAlgorithm( | ||
| conqueralg = ClA.ColCutGenConquer( | ||
| stages = [ClA.ColumnGeneration( | ||
| pricing_prob_solve_alg = ClA.SolveIpForm( | ||
| optimizer_id = 1 | ||
| )) | ||
| ], | ||
| primal_heuristics = [], | ||
| node_finalizer = ClA.NodeFinalizer( | ||
| EnumerativeFinalizer(optimizer = call_enumerative_finalizer), | ||
| 1, 0, "Enumerative" | ||
| ) | ||
| ), | ||
| maxnumnodes = heuristic_finalizer ? 10000 : 1 | ||
| ) | ||
| ) | ||
| ) | ||
|
|
||
| model, x, y, dec, B = build_toy_model(coluna) | ||
|
|
||
| function enumerative_pricing(cbdata) | ||
| # Get the reduced costs of the original variables | ||
| I = [1, 2, 3] | ||
| b = BlockDecomposition.callback_spid(cbdata, model) | ||
| rc_y = BD.callback_reduced_cost(cbdata, y[b]) | ||
| rc_x = [BD.callback_reduced_cost(cbdata, x[b, i]) for i in I] | ||
|
|
||
| # check all possible solutions | ||
| sols = [[1], [2], [3], [1, 2], [1, 3], [2, 3]] | ||
| best_s = Int[] | ||
| best_rc = Inf | ||
| for s in sols | ||
| rc_s = rc_y + sum(rc_x[i] for i in s) | ||
| if rc_s < best_rc | ||
| best_rc = rc_s | ||
| best_s = s | ||
| end | ||
| end | ||
|
|
||
| # build the best one and submit | ||
| solcost = best_rc | ||
| solvars = JuMP.VariableRef[] | ||
| solvarvals = Float64[] | ||
| for i in best_s | ||
| push!(solvars, x[b, i]) | ||
| push!(solvarvals, 1.0) | ||
| end | ||
| push!(solvars, y[b]) | ||
| push!(solvarvals, 1.0) | ||
|
|
||
| # Submit the solution | ||
| MOI.submit( | ||
| model, BD.PricingSolution(cbdata), solcost, solvars, solvarvals | ||
| ) | ||
| return | ||
| end | ||
| subproblems = BD.getsubproblems(dec) | ||
| BD.specify!.( | ||
| subproblems, lower_multiplicity = 0, upper_multiplicity = 3, | ||
| solver = enumerative_pricing | ||
| ) | ||
|
|
||
| function enumerative_finalizer(masterform, cbdata) | ||
| # Get the reduced costs of the original variables | ||
| I = [1, 2, 3] | ||
| b = BlockDecomposition.callback_spid(cbdata, model) | ||
| rc_y = BD.callback_reduced_cost(cbdata, y[b]) | ||
| rc_x = [BD.callback_reduced_cost(cbdata, x[b, i]) for i in I] | ||
| @test (rc_y, rc_x) == (1.0, [-0.5, -0.5, -0.5]) | ||
|
|
||
| # Add the columns that are possibly missing for the solution [[1], [2,3]] in the master problem | ||
| # [1] | ||
| opt = JuMP.backend(model) | ||
| vars = [y[b], x[b, 1]] | ||
| varids = [CL._get_orig_varid_in_form(opt, cbdata.form, v) for v in JuMP.index.(vars)] | ||
| push!(varids, cbdata.form.duty_data.setup_var) | ||
| sol = ClMP.PrimalSolution(cbdata.form, varids, [1.0, 1.0, 1.0], 1.0, CL.FEASIBLE_SOL) | ||
| var_was_inserted, sol_id = ClMP.setprimalsol!(cbdata.form, sol) | ||
| if var_was_inserted | ||
| mc_1 = ClMP.setcol_from_sp_primalsol!( | ||
| masterform, cbdata.form, sol_id, string("MC_", ClA.getsortuid(sol_id)), ClMP.MasterCol | ||
| ) | ||
| else | ||
| mc_1 = ClMP.getvar(masterform, sol_id) | ||
| end | ||
| # [2, 3] | ||
| vars = [y[b], x[b, 2], x[b, 3]] | ||
| varids = [CL._get_orig_varid_in_form(opt, cbdata.form, v) for v in JuMP.index.(vars)] | ||
| push!(varids, cbdata.form.duty_data.setup_var) | ||
| sol = ClMP.PrimalSolution(cbdata.form, varids, [1.0, 1.0, 1.0, 1.0], 1.0, CL.FEASIBLE_SOL) | ||
| var_was_inserted, sol_id = ClMP.setprimalsol!(cbdata.form, sol) | ||
| if var_was_inserted | ||
| mc_2_3 = ClMP.setcol_from_sp_primalsol!( | ||
| masterform, cbdata.form, sol_id, string("MC_", ClA.getsortuid(sol_id)), ClMP.MasterCol | ||
| ) | ||
| else | ||
| mc_2_3 = ClMP.getvar(masterform, sol_id) | ||
| end | ||
|
|
||
| # add the solution to the master problem | ||
| varids = [ClMP.getid(mc_1), ClMP.getid(mc_2_3)] | ||
| primal_sol = ClMP.PrimalSolution(masterform, varids, [1.0, 1.0], 2.0, CL.FEASIBLE_SOL) | ||
| return !heuristic_finalizer, primal_sol | ||
| end | ||
|
|
||
| JuMP.optimize!(model) | ||
| @show JuMP.objective_value(model) | ||
| @test JuMP.termination_status(model) == MOI.OPTIMAL | ||
| for b in B | ||
| sets = BD.getsolutions(model, b) | ||
| for s in sets | ||
| @test BD.value(s) == 1.0 # value of the master column variable | ||
| @test BD.value(s, x[b, 1]) != BD.value(s, x[b, 2]) # only x[1,1] in its set | ||
| @test BD.value(s, x[b, 1]) != BD.value(s, x[b, 3]) # only x[1,1] in its set | ||
| @test BD.value(s, x[b, 2]) == BD.value(s, x[b, 3]) # x[1,2] and x[1,3] in the same set | ||
| end | ||
| end | ||
| end | ||
|
|
||
| end |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters