System Module Active

Pre-Berth Optimization

Minimize weighted admission delay through globally aware berth assignment. This layer determines when and where each vessel begins service.

Layer Scope

Admission Scheduling Under Scarce Resources

Pre-Berth controls the time segment between vessel readiness and berth service start. It makes three key decisions that shape the entire downstream workflow.

Berth Assignment

Which vessel should use which berth, considering physical feasibility, yard proximity, and crane accessibility.

Service Start Timing

When berth service should begin for each vessel, balancing delay minimization against plan stability.

Admission Sequencing

How vessels should be ordered at each berth to minimize total weighted delay and downstream disruption.

Decision Variables

--x_{v,b} in {0,1}: vessel v assigned to berth b
--s_v: berth service start time for vessel v
--u_{i,j,b} in {0,1}: sequencing indicator -- vessel i before j on berth b

Hard Constraints

--Vessel readiness: s_v >= r_v (cannot start before ready)
--One berth assignment: each vessel assigned to exactly one berth
--Berth feasibility: infeasible vessel-berth pairs forbidden
--Same-berth non-overlap: no temporal overlap at a single berth
--Marine admission: pilot, tug, tide, channel constraints respected

Lexicographic Objective

--Level 1: min SUM w_v (s_v - r_v) -- primary weighted admission delay
--Level 2: min SUM C^plan_v -- plan stability / churn cost
--Level 3: min SUM C^risk_v -- fragility / robustness cost
--Secondary: min SUM C^priority_regret_v -- sequencing regret

Ontology Inputs / Outputs

--Graph nodes: SHIP, VOYAGE, BERTH, STS_CRANE
--Key edge: VOYAGE -> DOCKS_AT -> BERTH
--Input: vessel readiness times, priority weights, feasibility matrix
--Output: BerthPlan (vessel_id, berth_id, planned_berth_start, ...)
--Receives downstream feedback from Service Layer

Ontology Footprint

Graph Nodes

SHIPVOYAGEBERTHSTS_CRANE

Key Edge

VOYAGE → DOCKS_AT → BERTH

Objective Function

Lexicographic Admission Optimization

A multi-level objective that first minimizes weighted delay, then preserves plan stability, then maximizes robustness under uncertainty.

Primary Objective

\min \sum_{v \in \mathcal{V}} w_v \big( s_v - r_v \big)

where:
  w_v  = priority weight of vessel v
  s_v  = berth service start time
  r_v  = readiness time (fixed)

Since r_v is fixed, equivalent to:
  min  SUM  w_v * s_v

Minimize the total weighted admission delay across all vessels. Higher-priority vessels carry more weight, ensuring their delays are penalized proportionally.

Full Objective With Secondary Terms

\min \sum_v w_v(s_v - r_v) + \lambda_1 \sum_v C^{\text{plan}}_v + \lambda_2 \sum_v C^{\text{risk}}_v + \lambda_3 \sum_v C^{\text{regret}}_v

where:
  C^plan   = replanning / churn cost
  C^risk   = fragility cost (slack-poor plans)
  C^regret = bad sequencing regret

The full objective augments primary delay minimization with plan stability, robustness, and sequencing quality penalties weighted by tunable lambda coefficients.

Lexicographic Hierarchy

\text{Level 1: } \min \sum_v w_v(s_v - r_v) \quad\quad \text{Level 2: } \min \sum_v C^{\text{plan}}_v \quad\quad \text{Level 3: } \min \sum_v C^{\text{risk}}_v

Level 1 (primary):  Minimize weighted admission delay
Level 2 (secondary):  Minimize plan churn / instability
Level 3 (tertiary):  Minimize fragility under ETA uncertainty

Each level is solved strictly in priority order.
Lower levels cannot degrade higher-level solutions.

The lexicographic structure ensures delay minimization is never sacrificed for stability, and stability is never sacrificed for robustness.

Pre-Berth Quick Reference

Cadence

Runs before vessel admission. Re-triggered on ETA shifts, priority changes, or downstream feedback.

Key Variables

w_v -- vessel priority weight
s_v -- berth service start time
r_v -- vessel readiness time
C^plan -- replanning churn cost
C^risk -- fragility / robustness cost

Graph Nodes

SHIPVOYAGEBERTHSTS_CRANE

Key Output

// BerthPlan schema
vessel_id
berth_id
planned_berth_start
priority_weight
estimated_work_profile
marine_access_window

Recommended Solver

MILP, CP-SAT, or decomposition-based admission scheduler

IMPORTANT

Pre-Berth is structurally upstream (earlier in the decision chain) but NOT necessarily more important. The Service Layer is often more performance-sensitive. Pre-Berth sets the rhythm; Service Layer determines execution quality.

Constraints

Hard & Soft Boundaries

Hard constraints define physical and operational feasibility. Soft constraints guide solution quality toward globally efficient admission patterns.

Hard Constraints

Vessel Readiness

s_v >= r_v -- berth service cannot begin before the vessel is ready. Readiness depends on arrival, pilot availability, and tug allocation.

One Berth Assignment

SUM_b x_{v,b} = 1 -- every vessel must be assigned to exactly one berth. No vessel left unassigned, no double-booking.

Berth Feasibility

x_{v,b} = 0 if (v,b) not in F -- infeasible vessel-berth pairs are forbidden. Feasibility considers LOA, draft, cargo type, and equipment reach.

Same-Berth Non-Overlap

Two vessels assigned to the same berth cannot have overlapping service windows. Mandatory mooring/unmooring buffers included.

Marine Admission Feasibility

Pilot, tug, tide, and channel constraints must all be satisfied for the planned start time. No vessel begins service without valid marine access.

Soft Constraints

Yard Proximity Preference

Prefer berth assignments that reduce AGV travel distance to the primary yard blocks for that voyage's containers.

Crane Contention Avoidance

Avoid scheduling adjacent vessels in ways that create peak contention for the shared STS crane pool.

Downstream Impact Awareness

Factor expected STS inefficiency, AGV congestion, and yard-crane pressure from Service Layer feedback into admission ordering.

Priority Sequencing

High-priority vessels (contractual deadlines, alliance commitments) should be sequenced earlier when hard constraints permit.

Global Awareness

Downstream Feedback Integration

Pre-Berth is not a blind local optimizer. It receives structured feedback from the Service Layer to make globally informed admission decisions.

The Service Layer provides estimates of downstream cost under each candidate admission plan. Pre-Berth uses these signals to avoid plans that look locally optimal but create systemic congestion or fragility.

Expected Service Duration

How long the Service Layer estimates quay operations will take under the candidate berth plan.

STS Inefficiency

Projected crane idle time and interference cost from the proposed vessel-berth-time arrangement.

AGV / Yard Congestion

Downstream estimates of horizontal transport and yard-crane pressure given the admission pattern.

Robustness Risk

How fragile the plan is under ETA uncertainty -- plans with thin slack buffers score higher risk.

Future Berth Flexibility

Impact of the current plan on berth availability for vessels arriving in the next planning horizon.

Design Principle

Not a Blind Local Optimizer

Pre-Berth receives structured downstream signals and uses them to adjust berth assignment and sequencing decisions. A locally minimal-delay plan that creates severe crane contention or yard congestion will be penalized.

Signal Flow

Service Layer → Pre-Berth feedback loop. Signals include expected durations, inefficiency metrics, congestion forecasts, and robustness assessments.

Output to Service

BerthPlan: vessel_id, berth_id, planned_berth_start, priority_weight, estimated_work_profile, marine_access_window

IMPORTANT

Pre-Berth is a globally aware admission optimizer that receives downstream signals. It is NOT a standalone assignment engine. Every admission decision accounts for expected service duration, crane efficiency, transport congestion, and robustness under ETA uncertainty.

Execution Policy

Rolling Replan Triggers

Pre-Berth replans only when measurable thresholds are breached. The rolling-horizon approach preserves committed plans while adapting to new information.

Trigger

ETA shift exceeds admission slack buffer

Action

Re-solve berth assignment for impacted vessel and adjacent berth occupants. Preserve unaffected assignments.

Trigger

Marine resource unavailability (pilot, tug, tide window missed)

Action

Shift planned start time forward to next valid marine access window. Cascade downstream only if non-overlap is violated.

Trigger

Service Layer reports significant inefficiency under current plan

Action

Evaluate alternative berth assignments or sequence swaps that reduce projected downstream cost.

Trigger

Priority vessel arrives with urgent contractual deadline

Action

Inject priority vessel into sequence with bounded disruption. Stability penalty applies to displaced vessels.

Replan Philosophy

Pre-Berth uses a rolling-horizon approach. Only the impacted portion of the plan is re-optimized; committed assignments are preserved unless thresholds are breached.

Scope: Smallest feasible replan scope. Re-optimize only the affected vessel(s) and their berth neighbors.

Stability: Plan-change cost (C^plan) is an explicit lexicographic objective (Level 2).

Cascading: Downstream layers (Service, Yard, Gate) adapt to the new plan within their own cadences.

Warm-start: Previous solution used as warm-start for the solver. Full cold-start re-optimization is avoided.

KEY INSIGHT

The Pre-Berth layer is the stable admission backbone. It replans deliberately and only when risk signals justify the cost of change. Plan stability is not a nice-to-have -- it is an explicit optimization objective.