Decision Variables are the unknowns in an optimization problem whose values must be determined to optimize the objective function while satisfying constraints. They represent controllable decisions such as: production quantities, resource allocations, assignment choices, routing decisions, or timing. Variables can be: continuous (real-valued), integer (discrete), or binary (0/1 for yes/no decisions). Proper variable definition is critical for effective modeling - variables should capture all decision freedom while avoiding redundancy. Considerations: dimensionality, domains, symmetry, and implied constraints. Preprocessing techniques like variable fixing, bound tightening, and aggregation can reduce problem size and improve computational performance.