Public Services
Service sector · decision type · thirteen models
Operations research carries special moral weight when applied to public services — every misallocated fire truck, misassigned school seat, or unclaimed kidney affects lives directly. From RAND's fire-department deployment analysis for New York City (Walker, Chaiken & Ignall, 1974) to the market-design tradition of Roth, Sönmez & Ünver recognised by the 2012 Nobel Prize in Economic Sciences, this is the OR subfield where efficiency must be weighed against equity, accessibility, and institutional legitimacy. Thirteen applications below, organised along a service sector × decision type matrix anchored on Larson & Odoni's Urban Operations Research.
Why public-services OR matters
Three landmark moments · each verifiable, each consequential
Decision framework
Four lenses on the same thirteen applications
The primary taxonomy draws on Larson & Odoni's Urban Operations Research (1981, 2007 reprint) and the broader public-sector OR literature (Pollock, Rothkopf & Barnett, 1994). Applications are organised along two axes: the service sector (rows) and the decision type (columns). Dashed cells are honest gaps — decision problems that exist in practice but are not yet modelled here.
Icon flags an application with substantial critical-scholarship literature (see the Critical Lens tab). NEW marks applications added in the current restructure.
Public-services OR is defined by the tension between efficiency (minimise cost, travel time, makespan) and equity (ensure fair access, bounded worst-case outcomes, respect for affected communities). The plane below positions each application by its dominant objective. Hover a dot to read the full name; click to open the page. See Marsh & Schilling (1994) for equity measures in facility location and Karsu & Morton (2015) for an OR-wide survey of inequity-averse optimisation.
Public-services OR does not serve a single decision-maker. Models are commissioned by governments, implemented by nonprofits, and bear on the lives of citizens and communities — often with competing interests. The columns below group the thirteen applications by whose question the model most directly answers. An application can legitimately appear in more than one column; only the primary decision-maker is shown here.
This lens highlights applications whose mathematical formulations are in tension with documented real-world harms, reform histories, or impossibility results from the fairness literature. The individual pages cite peer-reviewed critiques. Silence on these critiques would be a doctoral-level failure of intellectual honesty; naming them is the minimum.
Chouldechova (2017). Fair prediction with disparate impact. Big Data.
Barocas, Hardt & Narayanan (ongoing). Fairness and Machine Learning.
See police-patrol.html.
Mehrotra, Johnson & Nemhauser (1998). Political districting MILP. Operations Research.
Ricca, Scozzari & Simeone (2013). Political districting review. Annals of OR.
See political-districting.html.
Abdulkadiroğlu, Pathak, Roth (2009). Strategy-proofness vs. efficiency in matching with indifferences: Redesigning the NYC high-school match. AER.
See school-choice-mechanism.html.
Application catalog
All thirteen pages · click a card to open the application
Current research frontiers
Where public-services OR is actively evolving
Algorithmic fairness in public decisions
Impossibility results (Chouldechova, 2017; Kleinberg, Mullainathan & Raghavan, 2017) show that calibration, predictive parity, and equalised odds cannot in general be satisfied simultaneously — forcing public-sector model designers to choose explicitly which fairness axiom to prioritise.
Market design for scarce public goods
Expanding the deferred-acceptance and kidney-exchange toolkit to refugee resettlement (Jones & Teytelboym, 2017; Bansak et al., 2018), organ allocation across other organs, daycare, and housing assignment. Each domain raises new incentive-compatibility and stakeholder-consent questions.
Humanitarian and climate-driven operations
Pre-positioning of relief supplies, shelter siting under sea-level-rise scenarios, and multi-stage stochastic planning for climate-driven displacement. The OR Society and INFORMS Humanitarian Operations communities treat this as the field's fastest-growing branch.
Equity-aware optimisation
Moving beyond min-max regret (Karsu & Morton, 2015) toward formal fair-division guarantees (Nash social welfare, envy-freeness, max-min share) for indivisible public goods — increasingly relevant for participatory budgeting and platform-mediated services.
Key references
Cited above · DOIs & permanent URLs