Court Hearing Scheduling
Job-Shop Scheduling · backlog · access to justice
A court has a weekly calendar: judges (who are like machines), courtrooms (another scarce resource), hearings that require specific judges and run for estimated durations, case-priority rules, and lawyer availability constraints. The scheduling problem is a close cousin of the classical job-shop scheduling problem. What makes it a public-services problem rather than a manufacturing one is that each delayed hearing carries human cost — a defendant in pretrial detention, a plaintiff waiting for a ruling, a family postponing resolution.
Problem context
Hearings, judges, calendars, backlog
US court systems face chronic backlog: in a typical state trial court, the time from filing to disposition for a civil case ranges from months to years; criminal cases are constitutionally bounded but still often approach those limits. Scheduling a hearing requires coordinating a judge, a courtroom, prosecuting/defending lawyers, witnesses, jury availability (for trials), and the defendant's or parties' presence — with deadline-driven priority rules (bail hearings within 48 hours, speedy-trial clocks, etc.).
Modelled as scheduling, it sits between job shop (hearings are jobs that require specific resources) and multi-mode resource-constrained scheduling (judges can substitute for each other for some hearing types but not others). Early OR court-scheduling work includes Hickman & Meier (1974); more recent applied operations research has addressed docket management, jury scheduling (Munger, Olson & Zimmerman 2022), and pretrial detention optimisation (Lawless et al. 2021).
Multi-stakeholder framing
Mathematical formulation
Job-shop / parallel-machine with due dates
| Symbol | Definition | Domain |
|---|---|---|
| J | Set of hearings to schedule | |J| = n |
| M | Set of judges × courtrooms (resources) | |M| = m |
| pj | Duration of hearing j | ℝ⁺ |
| dj | Due date of hearing j (statutory deadline) | ℝ⁺ |
| wj | Priority weight (case urgency) | ℝ⁺ |
| Ej ⊆ M | Eligible resources for hearing j (e.g. only civil judge) | |
| xjm | Binary: 1 if hearing j scheduled on resource m | {0, 1} |
| sj | Start time of hearing j | ℝ⁺ |
| Cj | Completion time of hearing j = sj + pj | ℝ⁺ |
| Tj | Tardiness = max(0, Cj − dj) | ℝ⁺₀ |
s.t. Σm ∈ Ej xjm = 1 ∀ j // each hearing scheduled on one eligible resource
on each resource m, no two hearings overlap in time // non-overlap machine constraint
Cj = sj + pj, Tj ≥ max(0, Cj − dj)
lawyer-attendance constraints: if hearings i, j share counsel, windows disjoint
xjm ∈ {0, 1}, sj ≥ 0
Complexity NP-hard (generalises weighted-tardiness job-shop scheduling, which is strongly NP-hard). Practical sizes (a few hundred hearings per week per court) solved by MILP + constraint programming or dispatching rules (Apparent Tardiness Cost / Weighted Shortest Processing Time).
Interactive solver
Dispatching-rule scheduler for weekly court calendar
Court Scheduler
Solution interpretation
Tardiness, throughput, and who waits
EDD prioritises hearings with near-term deadlines; good at reducing the count of tardy hearings. WSPT prioritises short, high-weight hearings; good at reducing weighted total tardiness when all hearings have similar deadlines.
ATC combines the two: the score for hearing j at time t is (wj / pj) · exp(-max(0, dj−t−pj) / κ¯p). This typically beats EDD and WSPT alone on weighted-tardiness objectives (Morton & Pentico, 1993).
FIFO is the baseline: process in order of arrival. It's procedurally fair in a naive sense but ignores deadlines and priority. Real courts run closer to EDD / FIFO hybrids; pure efficiency optimisation has deeper legitimacy concerns (see Limitations).
Limitations & Critique
Access to justice is not captured by makespan
Who waits matters, not just how much
A system that minimises total weighted tardiness can still systematically delay hearings for low-priority (= low-weight) parties. If “weight” correlates with case-type demographics, efficient scheduling can deepen disparities.
Pretrial detention is the hidden variable
A delayed criminal hearing for a detained defendant is not just a statistic; it's loss of liberty, job, housing. Weighted-tardiness objectives often ignore this qualitative difference unless explicit priority is given.
Counsel-availability asymmetry
Private counsel has scheduling flexibility; public defenders, often juggling 100+ cases, have near-zero flexibility. Naive JSP that treats all lawyer-constraints symmetrically systematically disadvantages public-defender cases.
Judicial discretion and continuances
Courts routinely grant continuances for legitimate reasons (witness unavailability, evidence delays). A scheduling system that assumes fixed durations ignores this fundamental source of stochasticity — real schedules degrade much faster than models predict.
Jury pool constraints
Jury trials require a randomly-selected, demographically representative pool that's unavailable during working hours for most jurors. Trial-day scheduling is at least as much a jury-availability problem as a judge-availability problem; many JSP models ignore this.
Trial / plea bargaining feedback
Faster hearings reduce the time discount on trial outcomes, which can shift plea-bargaining dynamics. A system designed purely to maximise throughput may inadvertently change who pleas and under what terms. The objective is not neutral to outcome.
Extensions & variants
Where court scheduling research is active
Stochastic hearing durations
Duration uncertainty leads to stochastic JSP / robust scheduling variants. Pinchevsky (2021).
Jury management
Munger, Olson & Zimmerman (2022) on jury-pool sizing and summoning optimisation under no-show rates.
Pretrial risk assessment
Algorithmic bail-decision systems (see police-patrol) — connected scheduling problem with COMPAS-style fairness issues.
Multi-court / regional scheduling
Cross-court resource sharing, assigned-judge pooling (cases can be heard in multiple counties).
Case flow analysis
Queueing theory applied to case arrivals and dispositions; Little's law diagnoses chronic backlog.
Priority / triage policies
Explicit priority structures with equity-aware weights (e.g. give extra weight to detained defendants).
Scheduling interpreters
Court interpreter availability is often the binding constraint for non-English proceedings; separate scheduling sub-problem.
Virtual / remote hearings
Post-COVID shift to virtual hearings changed scheduling flexibility; empirical work underway on outcome effects.
Key references
Cited above
Related applications
Scheduling & justice-adjacent problems