OCEL 2.0 Enrichment
DataSynth enriches OCEL 2.0 event logs with lifecycle state machines, multi-object correlation events, and resource pool modeling for realistic process mining datasets.
Lifecycle State Machines
Each document type follows a probabilistic state machine:
PurchaseOrder States
Created -> Approved -> PartiallyReceived -> FullyReceived -> Closed
SalesOrder States
Created -> Confirmed -> PartiallyDelivered -> FullyDelivered -> Invoiced -> Closed
VendorInvoice States
Received -> Verified -> Approved -> Scheduled -> Paid
State transitions include probabilistic branching (e.g., 10% rejection rate, 20% partial delivery).
Correlation Events
Multi-object correlation events link related process instances:
| Event | Objects Linked | Description |
|---|---|---|
| ThreeWayMatch | PO + GR + Invoice | Links purchase order, goods receipt, and vendor invoice |
| PaymentAllocation | Payment + Invoices | Links payment to one or more invoices |
| BankReconciliation | Statement + Payments | Links bank statement lines to internal payments |
Resource Pools
Resource pools model workload distribution across processors:
| Strategy | Description |
|---|---|
| RoundRobin | Cycle through resources equally |
| LeastBusy | Assign to resource with lowest current workload |
| SkillBased | Match resource skills to event requirements |
Enriched Event Fields
Each OCEL event includes:
from_state/to_state– lifecycle state transitionresource_id– assigned resource from poolresource_workload– current workload of assigned resourcecorrelation_id– links related events across objects
Configuration
ocpm:
enabled: true
lifecycle_state_machines:
enabled: true
resource_pools:
enabled: true
pool_size: 10
assignment_strategy: least_busy
correlation_events:
three_way_match: true
payment_allocation: true
bank_reconciliation: true
coverage_threshold: 0.95
Evaluation
The OCEL enrichment quality evaluator checks:
- State transition coverage (% of valid transitions observed)
- Correlation event completeness
- Resource utilization distribution