haul: Route Planner for EVE Online

I would like to share with you haul (Hyperspace Asset Unloading and Loading), a route optimization and trade automation tool I developed specifically for the virtual economy of the EVE Online. In this article, I'll walk you through the methodologies, algorithms, and technical implementations behind haul, highlighting key aspects such as route optimization, arbitrage detection, risk assessment, and performance optimizations I included in the project.

Route Optimization Algorithm

Key Factors for Optimization

When creating haul, I identified several critical factors that significantly impact successful asset transportation in EVE Online:

Warp Speed: This defines travel duration within solar systems, measured in astronomical units per second (AU/s). Faster warp speeds drastically reduce overall travel times.
Align Time: The delay between initiating warp and actual warp engagement. Ships with align times under two seconds reduce vulnerability, enhancing safety.
Cargo Capacity: This influences the maximum trade volume per journey, directly affecting profitability.
Risk Assessment: A thorough examination of solar system security ratings, historical attacks, and known hotspots to maintain safe travel routes.

Route Calculation Implementation

To ensure efficient navigation, haul uses Dijkstra's shortest path algorithm, enhanced with a custom heuristic I designed to integrate multiple factors into a cohesive cost metric:

\[ \text{TotalCost} = \text{WarpTime} + \text{AlignTime} + \text{RiskFactor} \]

Specifically, WarpTime is computed as:

\[ \text{WarpTime} = \frac{\text{Distance (AU)}}{\text{WarpSpeed (AU/s)}} + \text{JumpTime} \]

Here, JumpTime is a constant duration associated with stargate transitions (about 10 seconds per jump).

I also dynamically factored in risk assessments by regularly updating risk data from publicly available kill reports, mathematically represented as:

\[ \text{RiskFactor} = \text{KillDensity} \times \text{CargoValueFactor} \]

This allowed haul to adapt to changing player behaviors and emerging threats in real-time.

Data Acquisition and Processing

Real-time Data Integration

I designed haul to rely heavily on real-time data sourced from EVE Online’s official EVE Swagger Interface (ESI) API, including:

Market data (real-time buy/sell orders)
Player locations
Detailed solar system attributes (security status, system connections)
Kill reports for ongoing risk mitigation

Arbitrage Detection Approach

To accurately identify profitable trade opportunities, haul applies a structured process for analyzing market data:

Data Filtering: Initially filtering out items exceeding cargo capacity limits (800 m³) or falling below profitability thresholds.
Profitability and ROI Calculations:

I calculated profitability per cubic meter (profit/m³) as:

\[ \text{Profitability} = \frac{\text{SellPrice} - \text{BuyPrice} - \text{TransportCost}}{\text{CargoVolume}} \]

Return on Investment (ROI), measuring profitability against initial investment capital, was calculated as:

\[ \text{ROI} = \frac{\text{Profit}}{\text{InitialCapital}} \]

The results were then ranked by both profitability and ROI to produce optimized trade itineraries and routes.

Technical Implementation Details

Backend Development

I developed the backend infrastructure of haul using Python and FastAPI to handle API interactions efficiently. Key features include:

Asynchronous API processing via asyncio for concurrent data handling and reduced latency.
Real-time data streaming enabled by FastAPI’s StreamingResponse to provide continuous updates.

Frontend Development

For the frontend, I chose React for its ability to provide users with intuitive, real-time visualizations. Server-sent events (SSE) were leveraged to maintain synchronization between frontend and backend data, dynamically presenting optimal routes, risk assessments, and profitable trading opportunities.

Performance Optimizations

To ensure responsiveness and efficiency, I precomputed station-to-station paths, significantly reducing runtime computation loads. Additionally, I experimented with rewriting parts of the backend in Rust, which further enhanced performance, particularly in intensive route calculations.

Lessons Learned and Future Plans

Challenges Encountered

Triangle Inequality Violation: I faced challenges in accurately visualizing routes due to warp mechanics that sometimes violated expected mathematical principles, requiring careful handling.
Real-time Risk Data Management: Currently, risk data updates manually, which limits responsiveness to rapidly changing threats.

Planned Enhancements

Automating real-time risk assessment updates.
Developing a dynamic routing mechanism responsive to real-time market conditions and risks.
Fully integrating the experimental Rust backend for improved scalability.

Creating haul allowed me to combine algorithmic optimization, economic modeling, and software development into a cohesive tool tailored to the unique challenges of EVE Online. Throughout this project, I developed and refined critical skills in algorithm optimization, API integration, real-time data processing, and asynchronous programming, resulting in a robust and highly functional addition to my technical portfolio.