Mostafa Dadkhah Mostafa Dadkhah

A hybrid discrete-event and agent-based model of an underground haulage network. Each truck runs a custom statechart over capacity-constrained roads, passing bays and terminal queues. A custom routing model runs Dijkstra search on a dynamically masked graph that excludes occupied, reserved and blocked segments.

On this platform I built a short-interval control (SIC) system for conflict management. It predicts segment-occupancy conflicts and coordinates many trucks and simultaneous conflicts through right-of-way arbitration, node and bay reservations and rerouting. It resolves cyclic dependencies, so deadlocks and cycles of control failure do not occur.

• 4–16trucks, fleet sizes tested
• 48 × 30settings × replications
• ~9trucks before the depot becomes the bottleneck
• 0deadlocks in the reported campaign
• −27%delay per cycle vs. loaded-priority control
• ²⁄₃less detour distance than local-information policies

• Discrete-event simulation
• Agent-based modelling
• Statechart design
• Graph routing
• Fleet sizing
• Bottleneck analysis
• Design of experiments
• Replication analysis
• Parameter tuning
• Calibration
• Policy benchmarking
• Conflict detection
• Deadlock prevention
• KPI design
• Model verification

A rolling-horizon dispatcher copies the current simulated fleet state, plays several candidate assignment sequences forward, and commits the one with the least delay. It runs on top of the SIC traffic controller, so its gains add to the traffic-control gains.

Event logs and KPI accounting separate travel conflicts from service queues, and consistency checks tie delay components to cycle times, throughput and crusher utilization.

• 18–33%further delay reduction, on top of SIC
• +4–9%more completed cycles
• 700replications per scenario, four scenarios
• 63–75%of baseline delay traced to traffic

• Algorithm design
• Rolling-horizon optimization
• Simulation-based optimization
• What-if analysis
• Scenario analysis
• Event logging
• KPI accounting
• Throughput & utilization analysis
• Benchmarking
• Statistical comparison

A six-step design method that links operational capabilities to module responsibilities, interface contracts, state requirements and verification criteria. I used it to specify a four-level control architecture and applied it to the traffic-control and dispatch modules. Presented at IISE 2024 and CIM CONNECT 2024.

1. Operational capabilities
2. Module responsibilities
3. Interface contracts
4. State requirements
5. Verification criteria

• 6steps in the design method
• 4control levels specified

• Systems architecture
• Requirements engineering
• Interface specification
• Digital-twin design
• Control hierarchy design
• Verification planning
• Interoperability
• MQTT messaging
• API integration
• Technical communication