Conference Agenda

Nowadays, travelers can choose from an increasing variety of mobility services to reach their desired destination. Manually researching and comparing various offers involves a great deal of personal effort, which is why suitable decision support is required. The aim is to suggest a personalized selection of routes to travelers, whereby the selection strategy should consider several aspects: Depending on the personal weighting of criteria such as, e.g., arrival time, distance, costs, emissions, or number of transfers, different people may prefer different options. Furthermore, due to the multimodal nature of the problem, where various transportation options can lead to equal objective function values, it is relevant to include the available variety of transportation combinations.

The problem is formulated as a many-objective scenario-based model, for which at first the Pareto set for up to five preferences is determined using an approximation approach. Afterwards, a flexible, traveler-friendly number of Pareto-optimal solutions is selected, which adequately represent the variety of available offers by taking into account the diversity found in the objective as well as the decision space. The selection process is guided by travelers' self-assessed preference weightings, which, according to studies, do not accurately reflect their real-life actions and can only be considered as an indication.

The new selection strategy is tested on the multimodal network of the city of Hildesheim. Preliminary results show the effective reduction of the Pareto set to a suitable number of solutions, which concisely informs the traveler about reasonable alternatives and enables him to make an appropriate choice.

To cope with the fluctuating electricity generation by renewable energy sources and decarbonize the building sector, flexible electrical loads like electric vehicles or electric heating devices are necessary. As the building sector is a major emitter of greenhouse gas emissions, with the majority caused by the provision of heat, there is an increased need for optimal control of heating systems in buildings. In this paper, multi-objective optimization problems for the control of heating systems and electric vehicles are defined for typical German residential districts. Since different, partly conflicting objectives arise in residential areas (energy costs, CO2-emissions, peak loads grid, thermal comfort, etc.), we use approaches from the field of multi-objective optimization.

Next to exact methods for solving multi-objective optimization problems like the dichotomic method, we introduce a novel pareto local search method for controlling heat pumps. The goal is to minimize the energy costs and to maximize the inhabitant’s thermal comfort while not creating additional peak loads in the grid. The local search approach starts from a naive solution and iteratively tries to improve the solutions by searching for better feasible heating schedules in the neighborhood of the original solution.

We compare the approach to metaheuristics for multi-objective optimization like the evolutionary algorithm NSGA-II and to conventional control strategies, like the hysteresis control.

In light of the ongoing developments in the climate crisis, it is necessary to consider factors beyond the sole economic perspective in energy supply network planning. In this talk, we propose a modeling framework allowing for a three-pronged approach: besides minimizing the network costs, we additionally aim for lowering the network’s carbon emissions as well as the homeowner buy-in for energy-efficiency retrofit measures. While the former two are widespread aspects when optimizing energy supply networks, the latter is a rather untouched one – yet, a successful transformation towards a low carbon energy supply crucially depends on the homeowner acceptance. Following a short introduction of the

underlying model, we introduce the notions necessary for integrating homeowner accpetance yielding a multi-objective optimization problem. We conclude the talk by highlighting some mathematical aspects together with presenting some numerical results.