Module org.maxicp
Package org.maxicp.search.blackbox
package org.maxicp.search.blackbox
-
ClassDescriptionCommercial-solver-style black-box search orchestrator for modeling-level CP and scheduling problems.Configuration for exhaustive DFS/FDS phase.Strategy used to choose which decision variables are frozen in each LNS restart.Snapshot of an
IntervalVarin an incumbent solution.Configuration for LNS improvement phase.Configuration for restart-based feasibility phase.Logging granularity for phase execution.Conflict-driven adaptive branching strategy (dom/wdeg+Last-Conflict+ objective-polarity / phase-saving / solution-guided value selection + primary/secondary variable tiering).Selects which decision-variable indices are frozen for one LNS restart.Fragment selector that keeps variables from the same array together.PGLNS-inspired selector: keep low-impact vars fixed and relax high-impact ones.Convenience entry points for runningBlackBoxSearchfrom the modeling layer.Configuration object used by optimization solves.Configuration object used by feasibility solves.Generic, problem-independent structural analyzer for a constraint model.Legacy LNS fragment selection: each decision variable is frozen independently.Feasibility search phase based on repeated DFS restarts with persistent conflict learning.Large Neighborhood Search (LNS) improvement phase for scheduling and hybrid scheduling+integer models.Constructive and exhaustive scheduling search phase for models containingIntervalVars.