Technical Glossary
Comprehensive documentation for output parameters, computational logic, and AI archetecture.
Functions & Output Parameters
Total Cost
The final financial output representing the sum of all operational expenditures (Fuel + Maintenance + Tolls) across all calculated trips.
CO2 Emissions
Total carbon footprint of the mission. Calculated by multiplying distance and trips by the mode-specific emission factor (mT/km).
Mode Shift
The percentage of the supply chain that has transitioned from its original transport mode to a more optimized alternative.
Savings
The total capital retained by switching from the baseline scenario to the optimized route, displayed in both USD and Percentage.
Load Factor
The utilization efficiency of transport assets. Measures actual cargo mass against the maximum rated capacity of the selected mode.
Transit Time
The calculated duration from point A to B, factoring in mode-specific throttle speeds and historical port/hub dwell times.
Distance
The total route length in kilometers. Derived from dataset precision or synthesized via geodesic Haversine logic for 'gap' lanes.
Energy Intensity
Measures Joules consumed per ton-kilometer, used to rank the physical efficiency of different logistical topologies.
Resilience Score
A proprietary AI-derived metric (0-100) indicating the route's resistance to fuel spikes and port closures.
Optimization Algorithms
Deterministic Dijkstra Search
"Our core routing engine uses a modified Dijkstra's Algorithm to find the single most efficient path through the logistics graph."
- Weight Modes: Optimizes for 'Cost', 'Time', 'CO2', or 'Balanced' based on user priority.
- Dynamic Re-weighting: Every edge weight is recomputed on-the-fly to account for fuel price spikes.
- Edge Pruning: Automatically excludes nodes or routes marked as 'Closed' in scenarios.
- Computational Complexity: O(E + V log V), ensuring sub-50ms results even for 10,000+ nodes.
Yen's K-Shortest Paths
"To provide alternatives (Route 2, Route 3), we implement Yen's Algorithm to find the next-best paths without cycles."
- Strategic Deviation: Identifies spur nodes where a path could branch into a more resilient direction.
- Path Diversity: Ensures Alternative 1 and 2 are significantly different, not just minor variations.
- Ranking Model: Paths are ranked by their score delta relative to the baseline path.
- Sustainability Filtering: Used to find the 'Greenest' possible route even if it costs slightly more.
Haversine Geodesic Computation
"When datasets lack explicit edge distances, the system computes the 'as-the-crow-flies' distance using spherical trigonometry."
- Earth Radius Constant: 6,371 km used for precise global projections.
- Synthetic Air Injection: Used to create virtual flight paths between hubs lacking direct road links.
- Coordinate Mapping: Converts Latitude/Longitude pairs into a 3D vector space for distance measurement.
- Circuital Deviation: We apply a 1.2x multiplier to Haversine results for road transport to mimic real-world road curvature.
Hamming Mode-Shift Metric
"Measures the structural change between the baseline dataset and the optimized result."
- Comparison Vector: Compares transport modes at every leg of the supply chain.
- Shift Calculation: Shift % = (Number of Legs with Changed Mode / Total Legs) * 100.
- Insight Generation: High shift percentages indicate that the historical strategy was significantly sub-optimal.
- Multi-modal Detection: Automatically flags shifts from expensive Air/Road to efficient Rail/Sea.
Physics & Market Constants
Ground Level Constants
Mode Efficiency Matrix
| Mode | Mileage (KM/L) | CO2 (mT/KM) | Avg Speed |
|---|---|---|---|
| Truck (Heavy) | 3.8 | 0.120 | 60 km/h |
| Rail (Intermodal) | 12.0 | 0.040 | 45 km/h |
| Maritime (Vessel) | 45.0 | 0.015 | 25 km/h |
| Air (Cargo Jet) | 0.8 | 0.800 | 800 km/h |
Dynamic Fuel Usage
- • Distance
- • Mode Efficiency
- • Actual Load / Capacity
Our most advanced calculation. It penalizes fuel efficiency by up to 35% when the vehicle is running at 100% capacity (LoadFactor = 1.0).
Balanced Path Score
- • Calculated USD Cost
- • Transit Time
- • Fixed TimeValueFactor
The 'Secret Sauce' of our balanced optimization. It treats 1 hour of delay as being equal to $25 in lost operational value.
Advisor Neuro-Engine
Generative Reasoning Phases
Conversational Data Gathering
Using Gemini 2.0 Flash (Temp 0.7), the system interviews the user to understand constraints. It uses a structured prompt to track progress across 7 critical domains (Origins, Destinations, Volume, Cargo Type, etc.).
Synthetic Graph Synthesis
Once information is gathered, the **Neuro-Engine** switches to Temperature 0.3 (Precision Mode). It generates a compliant JSON graph including accurate Lat/Long coordinates and Haversine-calculated edges.
Semantic Explanation
Finally, the AI provides a natural language summary of why it chose specific paths, identifying strategic trade-offs between speed and resilience that pure math might miss.
Our prompts require the model to reason through fuel price impacts before emitting the final JSON structure.
OpenRouter gateway dynamic scaling ensures 99.9% uptime during simulation spikes.
The Stack
System Specifications
Technical Parameter Dictionary
LCO (Least Cost Path)
Calculated using Dijkstra's algorithm. It represents the global minimum of the cost function f(x) = ∑(LegCosts).
GHG Protocol Tier 1
Our CO2 math follows standard GHG reporting metrics for transportation fuel emissions factors (Scopes 1 and 3).
Delta Variance (δ)
The percentage of optimization. Formula: ((BaselineCost - ScenarioCost) / BaselineCost) * 100.
Bunker Surcharge
Specific to maritime legs. A variable cost adder tied to the 'Bunker' fuel index price index (1.0 = Base).
Edge Latency
The compute time required to traverse one adjacency list entry during graph exploration. Constant at <1μs.
Synthetic Route
A non-historical route generated by AI to fill gaps in legacy datasets, allowing 'what-if' modeling of new lanes.