CUBE ANALYST

Origin to Destination (OD) Matrix Estimation is the procedure of adjusting the input OD Matrices to match a set of target observed data.

Cube Analyst is the Cube Extension developed specifically for both static and dynamic estimating and updating of base year automobile, truck and public transit Origin-Destination trip tables. The Cube Analyst suite is comprised of two separate modules:

• Cube Analyst - The original matrix estimation program capable of a wide variety of static type estimation problems for private and public transport matrices.
  • Methodology based on Maximum Likelihood Method
  • Problem types: Small Static Matrices
• Cube Analyst Drive - The latest matrix estimation program designed for both static and dynamic estimation, with improved performance and advanced capabilities:
  • Methodology based on Data Assimilation Techniques
  • Problem Types: Static and Dynamic Matrices of any size

The matrix representing existing travel is one of the most valuable pieces of data in travel-demand forecasting. This matrix supports forecasting and almost all important comparative analyses. Cube Analyst enables planners to manipulate the extensive data set used to develop and update this matrix. Indeed, planners have successfully used Cube Analyst in various studies around the world.

Cube Analyst uses mathematical techniques to find trip matrices consistent with observed transport demand and count data. Cube Analyst reproduces hand-based methods more accurately and more efficiently.

Cube Analyst enables the user to exploit a wide variety of data that contributes to matrix updating and matrix development.

To use Cube Analyst, you supply observed travel-demand data like trip-end data collected in a shopping-center survey, traffic-count data organized into screen- and cut-lines, or movement or path data identifying travelers’ routes from origins to destinations or partial trip count data. Cube Analyst can use a wide range of low-cost, readily available data.

You can supplement this travel-demand data with quality weights. Quality weights provide tolerance bands for the data observations. Cube Analyst estimates the values that best fit the observations and their quality weights in a set of iterative calculations.

You can analyze the quality of the estimated matrix. Cube Analyst tools can characterize the extent of changes and help you find areas of significant change between input and estimated information.

Image 1 - Typical scatter plot and comparisons (TLD) between observed and modeled volumes

Within the Cube Base Application Manager interface, it is very easy to implement sophisticated procedures for the estimation and update of OD matrices, via the integration of feedback loops and convergence measures, as shown in the picture below.

Image 2 - Example of looping process for the estimation of dynamic OD matrices