|
Philip van der Wel Administrator Posts: 145
3/15/2019
|
Philip van der WelAdministrator Posts: 145
In our day-to-day customer support we find ourselves fielding similar questions from a range of users. The purpose of this help post is to consolidate the most frequently asked questions and, in doing so, create a valuable resource for our users and anyone who is considering using JunoViewer.
1. What types of condition data and distresses are used in JunoViewer?
Users can define their own distress parameters, tables, and columns to use in JunoViewer. Visual distresses such as potholes, bleeding, etc. can be imported and used, as well as high speed data such as roughness and rutting data recorded with a high speed profilometer. Our system can be easily configured to handle all of these data types using your own distress names and data types.
2. How is network condition assessed in JunoViewer?
JunoViewer offers many ways to display and report on present condition. In addition to strip maps, scatter plots, and trend graphs, you can create your own condition indexes and use these to summarise and plot condition. The framework can accommodate any measures users wish to utilise in assessing their condition data.
3. How are treatments assigned in JunoViewer?
Our modelling platform allows many different approaches to select the best treatment strategy. A popular approach is the use of Net Present Value or Incremental Benefit Cost Ratio of the treatment life-cycle cost, coupled with Optimisation. Similar approaches such as Maximum Benefit Cost Ratio can also be used. See this help post for more details on Benefit Cost Analysis (BCA) models.
Although these "value for money" approaches to modelling are popular, they require an accurate model to translate road condition to costs. In JunoViewer, this is often referred to as the "Cost of Use" parameter. In cases where clients do not have detailed data to calibrate a "Cost of Operation" parameter, or to otherwise support these theoretical/financial models, a simpler approach such as ranking based on condition can also be used effectively.
4. Does JunoViewer use Ranking or Optimization?
Please refer to the answer to the Question 3. Put simply, JunoViewer can do either Ranking (also referred to as Prioritization) or Optimization. You can pick a model type to best suit your network. Key considerations are: the scope of information you have available to drive and inform the modelling process.
If your network has relatively little detailed information, and specifically: if you do not have a well-calibrated model to translate condition to maintenance cost (referred to as "Cost of Use" in JunoViewer), then using a complex Optimization method based on financial indicators may not be a wise approach. In such cases, a robust Ranking approach using a well-chosen and balanced ranking parameter may be more effective.
5. Does JunoViewer use the HDM Models or can it incorporate HDM?
JunoViewer is a programmable modelling framework. As such, you can use any models from any source (like HDM). For example, if you want JunoViewer to model rutting based on the HDM4 model equation, then you simply need to program that equation into your DMS file (DMS means Deterioration Model Setup - you can find more information about it in this post).
Like HDM, JunoViewer can also do optimization based on Benefit Cost Analysis (BCA). JunoViewer BCA models can aim to optimize Net Present Value or Maximum Benefit Cost Ratio. JunoViewer does not presently offer the use of the Internal Rate of Return (IRR) for optimization analysis.
Also note that in the HDM documentation, there are three levels of model calibration. The most advanced method of calibration (Level 3) involves (a) improved data collection, and (b) fundamental research. At this level, alternative model functions can be developed based on network specific data. In the modern asset management era, there is often several years of data and local knowledge available that makes it possible to develop evidence-based, network specific models.
The HDM models suggest that Level 3 calibration is a "long-term endeavour". We at Lonrix believe this is an outdated concept. Many road networks now have several years of high speed data at 10 m spacings, coupled with detailed information related to surfacing ages and types etc. In a modern framework such as JunoViewer, it is easy to rapidly generate historical deterioration rates for available condition parameters.
For a network with several years of condition data within JunoViewer, it is possible, by using our SmartRate calculator, for an analyst to generate a full set of observed deterioration rates for their network within a day or two. This data can then be exported and analyzed in a statistical framework such as R. Using these modern tools, an experienced analyst should be able to develop evidence-based, network specific models within two to three weeks. It is almost certain that these models -developed for the network in question using actual data collected on that network - will be superior to more generic models developed using data spread over several countries.
Please also see details in this post regarding a default model that implements some of the HDM concepts in JunoViewer.
6. Does JunoViewer use a Looping Algorithm or a Single Year Ranking?
This question pertains mainly to the use of a Prioritization/Ranking algorithm. If you are using a Ranking algorithm, JunoViewer always performs a multi-year ranking based on your chosen ranking parameter. As such, it loops and ranks repeatedly over all analysis years. For each year, treatments are then applied in order of ranking for that year. The improvement in condition due to such applied treatments are taken into account in the ranking for the following years. This process is repeated for all years in the modelling period. Again, this applies only for a Ranking model. for Benefit Cost Analysis (BCA) models, a different algorithm is applied (see Question 7 below). For a broader perspective on modelling types, please refer to the answers to Questions 3 and 4 above.
7. What type of Optimization Approach does JunoViewer use?
For Benefit Cost Analysis (BCA) models, JunoViewer uses the concept of Pareto Optimality coupled with heuristic optimization to select an set of projects that will aim to optimize total benefit, or to maximize benefit cost ratio (i.e. leaning toward minimizing costs for a high, if not maximal, benefit). This is an area of continued research and we are continually exploring more effective algorithms. Among other avenues, we are currently exploring the use of Genetic Algorithms and may in future provide this as another option for optimization.
8. My Budget is not being Fully Utilized (Benefit-Cost Analysis Models)
As explained in this help post, when you run a BCA model, the model will generally only select projects that fall on, or close to, the Pareto Frontier. In some cases, depending on how your projects are spaced over years, and also on the available budget in different years, it may happen that the projects that lie on or close to the Pareto Frontier cannot be fitted into the available budget and more of the generated projects need to be considered. In such a case, you should consider increasing your Pareto Frontier Sensitivity, which you can do on the General Sheet of the DMS file. Please refer to the details for the Pareto Frontier Sensitivity as defined in this help post.
edited by admin on 10/13/2020
|