This topic covers all elements of Deterioration Modelling in JunoViewer Web
Grouped, Ranked, Cost-Constrained Model
|
Fritz Jooste Administrator Posts: 81
2/10/2020
|
Fritz JoosteAdministrator Posts: 81
The Grouped, Ranked, Cost-Constrained model is one of the work-horses of the JunoViewer modelling framework. It was one of the earliest models included in the framework and is invaluable for Multi-Carriageway Road networks or any other network where treatments in certain areas (e.g. adjacent lanes) need to be done at the same time.
Note that this model can be used with or without Key Performance Indicator (KPI) constraints. Details on how to use this model with KPI constraints are shown further down in this post.
The Grouped, Ranked, Cost-Constrained model uses the ranking/prioritization approach, but with the benefit of grouped ranking. With grouped ranking, you can assign each segment of your network to a specific group. Ranking is then performed on groups rather than on individual segments. The key outcome from this approach is that each group will be treated at the same time. Note that this does not mean that all segments in a group will receive the same treatment. Each segment will receive the treatment (if any) that is triggered for that segment when the group is treated.
For example, all lanes on a specific section of a multi-lane road can be assigned to a single group, as shown in the schematic below. In this figure, we have four groups: Section A1 and A2 in one direction, and Section B1 and B2 in the other direction. As you can see, each group contains 15 segments.
We want the model to assign treatments to each group at the same time. This ensures that adjacent lanes are treated at the same time. The ranking parameter used for ranking groups is specified on the General sheet of the DMS file. In particular, we use the General sheet parameter "Work Section Ranking Parameter" to determine which model parameter to rank on, and we use parameter Work Section Ranking Statistic to determine what statistic is used to combine rankings on individual segments within a group. For example, if the Work Section Ranking Statistic is set to the "Mean Value" then the ranking parameter for the group will be calculated as the average of the rankings of all individual segments. You can use the chosen statistic to determine whether ranking should be based on the worst segments in the group (e.g. using a higher percentile) or based on the general condition of the group (e.g. using Mean Value)
Say for example, the condition on Section A2 in the above figure is such that this group is ranked highest for treatment. Let's assume also that the budget or work section count constraints are such that only one group can be treated in a specific year. Then the model will only treat Section A2 in that year. Treatments assigned to individual segments will depend on segment-specific condition and your defined treatment triggers. For example, you may get an assignment that looks like this:
Note that not all segments in Section A2 received treatments. This may be because those segments are predicted to be in good condition at the time of treatment. Clearly, the assigned treatments are not all practical (some overlays only in one lane), but what the model does is give you an approximate estimate of the total treatment quantum and cost for that group in a certain year. Operationally, you will need to adjust the treatments to ensure surface levels in adjacent lanes are appropriate.
If you need the model to enforce certain treatments in adjacent lanes you have two options: (1) you can use the Grouped, Ranked, Cost-Constrained, Slow Lane Governs model in which treatments are only triggered on the slow lanes, with fast/middle lane treatments applied automatically based on which treatment was triggered in the slow lane. This model requires that you use the Traffic Model in which you need to specify lane codes considered as slow lanes; OR (2) you can talk to Lonrix about developing a customized model to suit your treatment types and specific needs.
edited by admin on 2/10/2020
|
|
|
0
link
|
