Optimization of Daily Activity Chains
The initial assumption was that some activities performed by the users during a day are not necessarily fixed temporally and spatially, therefore they can be carried out in different times or locations. The order of flexible demand points can be also changed. Before and after an activity a travel phase is realized by different transportation modes. By introducing flexible demand points, it is possible to find all combinations and to choose the optimal activity chain by implementing a solution for the TSP-TW problem.
The developed algorithm takes into consideration many constraints, as opening times of the shops or maximum waiting times before the planned arrival. During the implementation 3 different modes of transportation were determined: car, public transport and public transport with car-sharing opportunity. The optimization criterion was the minimum travel time, as the most important parameter. Also other parameters can be taken into account (e.g. comfort features), but these are generally hard to be quantified.
As an output of the optimization Pareto optimal results are presented, where the parameters are the number of postponed activities and the total travel times. The simulation of activity chain optimizations was performed on arbitrarily chosen test networks in Budapest using Matlab. In case of car usage about 8%, with public transport about 10% and with car-sharing opportunity about 14% decrease of the total travel time was realized. The elaborated method can be build in an advanced information service.
Domokos Esztergár-Kiss: Optimization of multimodal travel chains
PhD thesis, BME Dept. of Control for Transportation and Vehicle Systems, 2016.