30 November 2022
Ionic thermoelectrics: Mission Impossible?
Subject: Applied Physics
Lecturer: Dan Zhao Jonsson, ITN
Location: Campus Norrköping, K3, Kåkenhus
Facing the exacerbate energy crisis, thermoelectric materials are considered one of the most promising solutions to convert earth-available waste heat into electricity. However, the high raw materials cost and complicated manufacturing of the classic semi-conductive metal alloy-based thermoelectric generators are limiting large-scale application. On the other hand, electrolytes that can generate much higher thermal induced voltages are cheap, environmentally friendly and easy to process. In comparison to the extensive exploitation of electronic semi-conductive materials, electrolytes were missing in the scope of thermoelectrics for a long time. What are the challenges to apply electrolytes as ionic thermoelectric materials? How will they perform compared to electronic materials? Are we facing a target of “Mission Impossible”?
13 September 2022
Cost Partitioning for Classical Planning
Subject: Computer Science
Lecturer: Jendrik Seipp, Department of Computer and Information Science (IDA)
Location: Ada Lovelace, B Building, Campus Valla
Classical planning is the task of finding a sequence of actions that transforms a given initial situation into one that satisfies certain goal criteria. The dominant approach for solving classical planning tasks optimally is to run the A* search algorithm with an admissible heuristic, that is, a function that never overestimates the true distance to the nearest goal state. Usually, a single heuristic is unable to capture enough details of a given planning task. Instead, we can combine multiple heuristics in a way that preserves admissibility. The strongest approach to do so is called cost partitioning, where the cost of each action is divided among the heuristics. In this lecture, I will present the main algorithms for cost partitioning and compare them to each other. Then I will show how they can be used to obtain state-of-the-art optimal planners.