Assignment Overview

Key steps included in the the PTM Assignment, include;

1. Initialisation - The method initialises the assignment index and iterates over the assignments to be run.
2. Events Assignment - For each assignment, creates an EventsAssignment instance to handle event-specific details.
3. Base Demand Matrix - Retrieves the base demand matrix for the assignment.
4. Impedance Matrix - Depending on whether the model is elastic or not, obtains the appropriate impedance matrix.
5. Assignment Specification - Generates the assignment specification and results specification.
6. Multi-Class Processing - If multi-class assignment is required, processes the class and prepares the specifications.
7. Run Assignment - Runs the assignment using the appropriate method (e.g., congested, capacitated, or schedule assignment) based on the model type and crowding settings.
8. Elastic Assignments - If the model is elastic, it reinitializes the class list and runs the elastic assignment process, including recalculating project demand and running post-elastic assignments.

Frequency Model Assignments

Extended Transit Assignment

A multimodal, multipath transit assignment that generalizes the Standard Transit Assignment (optimal strategies) to allow for the the explicit modelling of origin-connector choice, sensitivity to travel time (in addition to headway) for choice of transit services at common stops, and more. Also provides saving of strategies and sophisticated post-assignment analysis capabilities.

Congested Extended Transit Assignment

A transit assignment that models congestion aboard vehicles using volume dependent cost functions. The dependence of the link cost on transit volume may represent a slowing down of the transit vehicle due to the number of passengers or a generalized cost including a discomfort term that increases as vehicles become crowded.

The travel cost function on segment a is given by a fixed (free flow) travel time ca0 and a volume dependent congestion function da (va) in the form ca(va) = ca0(1 + da (va)). In order to enable the transit time functions to reflect congestion costs, original transit time functions have to be multiplied with the term (1+usx), where usx is the congestion attribute (us1, us2 or us3). The tool modifies the functions temporarily to include this term unless the option Use current functions without modification is chosen, in which case congestion is taken into account only on segments for which the transit time function already contains the congestion term.

During the assignment steps, the user defined segment attribute usx will contain the value of the in-vehicle congestion term. It is a function of the V/C ratio, where the capacity of the transit line is computed as 60*vcapt/hdw. For PTM the following function is used; BPR— For BPR-type (default), the congestion term is computed as α * (V/C)β, where α is the weight of the congestion term and β is the exponent.

Capactitated Extended Transit Assignment

A transit assignment that uses the Method of Successive Averages (MSA) to obtain the transit flows that correspond to the congestion of transit segments. In-vehicle congestion functions (as in the Congested transit assignment tool) and increased waiting times at stops due to transit vehicle capacities are used to model the decreased comfort level in-vehicle as well as increased waiting times at stops

Two congestion functions are needed for this assignment;

• A segment congestion function describes how discomfort increases on-board crowded vehicles.
• A headway congestion function describes how waiting times at stops increase due to crowding. The segment congestion function is specified the same way as in the Congested transit assignment tool.

The increased waiting time is modelled through the use of effective frequency. The effective frequency of a line is the frequency which yields the waiting time obtained by the adjusted headway. It is modelled with a continuous function which increases as the residual capacity of the vehicle decreases. The increased waiting time at a stop may be modelled by using steady state queuing formulas, which take into account the residual vehicle capacity, the alightings and the boardings at stops. The segment-specific effective headway increases as the transit flow reaches capacity (but cannot exceed 999.98). The functional form used in the tool is: Min [ headway/(1-(boardings/residual capacity)^headway_adj),999.98]

The results are sensitive to the Headway adjustment exponent (headway_adj). In practice the Headway adjustment exponent may take a wide range of values and must be calibrated for a particular application. Smaller values tend to penalize boardings at demand levels well below capacity, whereas larger values will penalize boardings only as demand approaches close to capacity. In all cases when demand for a particular transit service at a particular stop exceeds the available capacity the effective headway will be set to the maximum (999.98).

Timetable Model Assignments

Schedule Based Transit Assignment

A complete schedule-based assignment which incorporates detailed itinerary and departure and arrival information in a time-dependent assignment in which travellers make explicit choices regarding early departures and late arrivals. It is useful for services which operate on irregular headways throughout a planning period for suburban and inter city applications such as national or regional services.

Crowded Schedule Based Transit Assignment

A schedule-based transit assignment that models congestion aboard vehicles using increased boarding costs at stops and in-vehicle congestion functions.

Two congestion functions are needed for this assignment;

• A segment congestion function describes how discomfort increases on-board crowded vehicles.
• A headway congestion function describes how waiting times at stops increase due to crowding.

The segment congestion function models the influence of a traveller’s path choice based on the discomfort experienced inside the vehicle. During each iteration of the crowded schedule-based transit assignment, the congestion term is added to the original in-vehicle cost as congestion_cost * transit_time, where transit_time includes the dwell_time. This congestion term is saved in segment extra attribute @ccost.

The functional form used in the tool is: BPR— For BPR-type (default), the congestion term is computed as α * (V/C)β, where α is the weight of the congestion term and β is the exponent.

The boarding congestion function models the influence on the traveller’s path choice based on the fact that the vehicle is already full at a boarding node. At each iteration of the crowded schedule-based transit assignment, the boarding cost congestion term is added to and saved in the segment extra attribute @bcost.

In a Schedule-based transit assignment, the waiting time of a traveller is the difference between the time at which the traveller arrives at a stop and the time when the vehicle he boards leaves the stop. In a Crowded schedule-based transit assignment, if the first vehicle the traveller wishes to board is full, the boarding penalty computed below would be too great, and the traveller opts to wait for another vehicle (or changes path in some other way). For clarity, if the traveller opts to wait for another vehicle, the computed waiting time would take into consideration the additional wait at the stop for the later, less costly vehicle. The following options are available to compute the boarding congestion term: Default— This option uses a capped BPR function, with the Congestion term weight and Congestion term exponent as specified. The boarding penalty is computed as follows when space = capacity − transit_volume is greater than 0: penalty = min(penalty, max(0, α × (transit_boardings / space) β)) where α is the weight of the congestion term and β is the exponent, and the initial penalty = 999999.