Rail Vehicle Optimization System Using Graph Network Flow
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optimizing vehicle utilization in freight railroads is challenging due to improper loading, leading to inefficiencies such as under-loaded or over-loaded vehicles, resulting in late deliveries and customer dissatisfaction.
Innovation Solution
An integrated system that includes a planning cockpit for repositioning empty vehicles, considering vehicle compatibility, tolerable delay, waiting cost, repositioning cost, and long stand cost, to optimize vehicle allocation and reduce operational costs, using an in-memory database for real-time analytics and optimizing vehicle routes and assignments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual vehicle assignment methods are used, then operational simplicity is maintained, but vehicle utilization efficiency deteriorates due to improper loading and suboptimal routing decisions
Solution Approach 1:
The patent replaces manual mechanical assignment processes with an automated computer-based optimization system that uses algorithms to determine optimal vehicle assignments, routes, and loading configurations, thereby improving utilization efficiency while managing complexity through automation
Solution Approach 2:
The system enables self-service optimization by automatically analyzing order requirements, vehicle availability, and routing options to generate optimal assignments without requiring manual intervention, allowing the system to serve itself in making optimization decisions
2Reliability
If optimized vehicle routing and assignment is implemented, then delivery timeliness is improved, but computational complexity and processing time increase
Solution Approach 1:
The system performs preliminary optimization calculations and generates vehicle assignment recommendations in advance of actual vehicle deployment, allowing time-consuming computations to be completed beforehand so that real-time decision-making is expedited without sacrificing optimization quality
Solution Approach 2:
The optimization system dynamically adjusts parameters such as vehicle capacity utilization thresholds, routing preferences, and loading configurations to balance computational efficiency with delivery timeliness, modifying these parameters based on operational conditions to reduce processing time while maintaining reliability
3Productivity
If comprehensive optimization considering multiple factors (loading, routing, timing) is applied, then overall operational efficiency is improved, but the difficulty of planning and execution increases
Solution Approach 1:
The patent implements a universal optimization platform that simultaneously handles multiple optimization objectives including vehicle assignment, routing, loading configuration, and timing coordination through a single integrated system, thereby improving overall operational efficiency while reducing the complexity of managing multiple separate planning processes
Data Source
AI summary
Various embodiments of systems and methods for optimizing vehicle utilization are described herein. The method includes determining demand, supply, and availability of vehicles. A graph with nodes (representing source and destination stations of orders) and arcs connecting nodes is generated based on validity conditions, e.g., there is network connectivity from source to destination node, etc. The arc is assigned for repositioning, waiting, and loaded move based upon input such as tolerable delay for orders, waiting cost at stations, repositioning cost, long stand cost, revenue generated from orders, order priority, etc. A network flow solver is successively executed to solve the graph to maximize profit, optimized utilization of vehicles, fulfill order on time, and balance demand and supply. Output tables including key performance indicators (KPIs) and information of fulfilled and unfulfilled orders are generated. Vehicle allocation, repositioning, waiting, and loaded move may in turn are optimized based on the output tables.