Overview
Goal programming models decisions with several aspiration levels instead of a single natural objective. The formulation introduces deviation variables and minimizes the unwanted shortfalls or excesses around each goal.
It fills the gap between plain linear programming and broader multi-objective optimization, especially in managerial settings where targets are easier to negotiate than abstract weights.
Core ideas
Aspiration levels
Aspiration levels are target values for goals such as budget, service, output, risk, or staffing balance.
Deviation variables
Deviation variables measure how far a solution falls below or rises above each target.
Preemptive priorities
Preemptive priorities optimize higher-priority goals before lower-priority goals are considered.
Weighted goals
Weighted goals combine deviations in a single objective when priorities are tradeable.
Soft constraints
Soft constraints allow controlled violations that are penalized rather than declared infeasible.
How to use it
- 1List the goals separately from the hard constraints.
- 2Define the aspiration level, direction, and units for each goal.
- 3Introduce positive and negative deviation variables where target misses matter.
- 4Choose whether goals are weighted together or ordered by priority.
- 5Review the achieved deviations with stakeholders rather than reporting only the objective value.
Applications
- Budgeting: balance cost limits against service, coverage, and equity targets.
- Workforce planning: trade off staffing cost, preference satisfaction, overtime, and coverage.
- Public policy: represent negotiated targets for access, fairness, cost, and emissions.
- Portfolio design: compare return, risk, liquidity, and allocation targets.
- Production planning: balance output, overtime, inventory, and service goals.
Common pitfalls
- Choosing weights or priorities without stakeholder agreement.
- Comparing deviations measured in incompatible units without normalization.
- Treating a soft goal as if it were a hard feasibility requirement.
- Hiding tradeoffs by reporting only the aggregate weighted objective.
Resources
- NEOS Guide — Multiobjective Optimization
Useful context for goal programming as a multi-objective modeling approach.
- MIT 15.053 Optimization Methods
Introductory optimization course with management-science modeling material.