Bus allocation to short-turning and interlining lines.

We propose injecting flexibility into public transport service planning by introducing a demand driven method for generating and assigning buses to short-turning and interlining services. This study formulates, solves and applies the problem of assigning vehicles to the lines of a bus network subject to the dual objective of (a) improving the passenger waiting times at stops and (b) reducing the operational costs. At first, the vehicle allocation problem is expanded with the explicit consideration of interlining and short-turning lines that provide greater operational flexibility. The paper introduces a rule-based approach for generating interlining and short-turning lines that are considered as "virtual lines" because some of them might remain inactive if their operation does not improve the vehicle allocation solution. The bus allocation problem to existing and virtual lines is modeled as a combinatorial, multi-objective optimization problem and is solved with a Genetic Algorithm (GA) meta-heuristic that can return improved solutions by avoiding the exhaustive exploration of a combinatorial solution space. The vehicle allocation to existing and virtual lines is applied to the bus network of The Hague with the use of Automated Fare Collection (AFC) data from 24 weekdays and General Transit Feed Specification (GTFS) data. Sensitivity analysis results demonstrate a significant reduction potential in passenger waiting time and operational costs without adding a large number of short-turning and interlining line options that could impede the practicality of the bus services