The procedure is essentially divided into a construction process for line generation and a subsequent selection and evaluation of line concepts.
In the first part, line routes are constructed that start from and end at the specified terminal stops. All routes courses are found that only use the links and turns of the specified transport system that can be traversed and do not involve too great a detour. The detour test ensures that the length of the route course between every two stops is not greater than a coefficient multiplied by the distance of the shortest path (regardless of permitted transport systems). The detour test takes into account the fact that obstacles such as rivers or buildings have to be avoided even if the route is ideal. The coefficient in the comparison between the constructed line and the ideal shortest path decreases with the length of the shortest path. In other words, larger detours can be taken locally, while the overall route course should follow a direct line as closely as possible.
For each line route found, an opposite direction is searched for in the set of constructed line routes or created if necessary. The respective line routes of the two opposite directions are combined into one line, so that a generated line always consists of two line routes (one in each direction). No ring lines are generated. If these are desired, they must be specified as an existing line in the procedure with a selectable headway.
When lines are created, all existing lines that are not marked as predefined lines by their attribute value are deleted.
The resulting line pool forms the basis for the next steps in the selection and evaluation of line concepts. To check this first step, a maximum network is output that contains all created and specified lines with the highest headway.
|
Note: It is worth checking the pool of lines created. Here you can check whether useful lines have been created before selecting and evaluating lines. Settings for detours, trafficability of links and turns, or selection of terminal stops can be adjusted until the generated line pool meets expectations. |
In the second part, line concepts are created and evaluated from the line pool by selecting lines and specifying a frequency. The assessment is based on the assignment of demand to the maximum network. A special assignment without parameters is used for this purpose. Based on the resulting volumes, line routes are selected that serve high direct distance passenger kilometers. Unserved demand is reduced by adding further lines. The allocation of a headway is based on capacity considerations that take into account the use of the line route on the one hand and the capacity of the vehicles used on the other. The headway is selected from a series that was previously specified in the parameters.
The selected line concept is now evaluated using a weighted objective function. The evaluation criteria must be standardized. This is done using the average evaluations of the individual objective function components of the start solutions.
The objective components used consist of components from the passenger's perspective and from the operator's perspective. This gives you the opportunity to find a balanced weighting of the competing requirements for a public transport service.
The following objective components are taken into account:
- Mean capacity violation: A capacity is assigned to the vehicles. The mean capacity violation is calculated from the sum of the travel times that passengers spend "standing" in vehicles, i.e. the passengers above the capacity of the vehicle are multiplied by the travel time that this condition lasts ("standing minutes"). There are no passengers left behind at the boarding stop, i.e. all passengers are always transported regardless of the volume.
- Mean additional number of transfers: The mean additional number of transfers is the difference between the average number of transfers on the paths of a relation in the evaluated supply and the average number of transfers on the paths of the maximum supply, multiplied by the demand of the relation and summed over all relations.
- Mean detour factor: The mean detour factor is the ratio by which the average travel time on the paths of a relation in the evaluated supply exceeds the average travel time on the paths of the maximum supply, multiplied by the demand of the relation and summed over all relations. Minimum and maximum service level: A lower and upper limit for the number of departures per hour can be set for each stop point. This component represents the sum of all exceedences and shortfalls in the network. It therefore equally measures customer expectations or the requirements of the local transport plan (especially lower limits) and/or operational constraints (especially upper limits).
- Number of lines above budget: The evaluation of the number of lines in the transport supply must be assessed, as a high number of lines enables many direct trips for the passenger on the one hand, but makes the supply untransparent. The user specifies an expected number of lines for the transport supply. The exceedence of this number is evaluated in the objective function. Selected lines whose line routes are completely contained in the line routes of another line are marked as reinforcement lines of the longer line routes. The reinforcement line is not included in the evaluation of the number of lines in the network.
- Number of vehicles above budget: The vehicle requirement is estimated from the link run time per direction, the specified layover time, and the headway of each line and added across all lines of the line concept. The indicator thus characterizes a component of operating and depreciation costs. Optionally, a budget of available vehicles can be specified for each transport system. Only the proportion of the number of vehicles that exceeds the budget is then assessed.
- Share of empty seat time: The total capacity kilometers and passenger kilometers per line can be calculated from the volumes of the time profiles on the one hand and the capacity of the vehicles and the headway on the other. The difference between the two is the empty seat kilometers, and these are set in relation to the total capacity kilometers. The higher this value is, the more inefficient the supply is, because many of the available seats remain unused over longer distances.
- Unserved passenger kilometers: The unserved passenger kilometers are calculated from the number of passengers on relations for which there is no supply with this line concept (or for which more than three transfers would be required), but the maximum supply offers a supply multiplied by the direct distance of the relation and added over all affected relations. Relations that already lose out to the competing mode in the maximum network or require more than three transfers do not play a role.
The selection and evaluation of the line concepts are part of a genetic algorithm. Various start solutions are generated. Good elements of the solutions are retained and others are varied. The solutions found gradually improve. The procedure ends if no better solution is found after a specified number of generations or if the maximum number of iterations is reached.
In addition to the line concept with the best evaluation, the procedure also provides a number of solutions that are almost as good. A line concept consists of the selected lines (line routes) and an attribute that contains the headway of the service. Vehicle journeys or associated departure times are not generated. On this basis, it is possible to immediately calculate a headway-based assignment that retrieves the headway information via the relation of time profile to line.
|
Note: The "Create regular timetable" add-in can generate vehicle journeys by specifying a time range and the headway information. The headway offset optimization (Headway offset optimization) sets optimal departure times to minimize wait times for passengers across the network. |
In order to assess, after the procedure has been carried out, by how much a solution could be improved in one dimension if a deterioration in another component were acceptable, all Pareto-optimal solutions found are also optionally output. The maximum network, i.e. the network that contains all the lines in the line pool, is also provided as a scenario for checks.
