Optimality and Space in Weakly Constrained Everyday Activities
- Petra Wenzl, Institute for Artificial Intelligence, University of Bremen, Bremen, Germany
- Holger Schultheis, Institute for Artificial Intelligence, University of Bremen, Bremen, Germany
AbstractThe action order of most everyday activities is only weakly constrained: When setting the table, for example, the order in which the items are placed on the table does not matter if all required items are on the table eventually. Little is known about how humans deal with weakly constrained sequences. Consistent with research on local optimality of human behavior and the "law of less work", we propose that the order of weakly constrained sequences is not chosen arbitrarily but due to preferences, with the overall goal to minimize cognitive and physical effort. We implement and validate a stepwise-optimal model for table setting, revealing ordering preferences based on distance, functional relations between items, and reachability. The model's success has implications concerning action organization in weakly constrained sequences as well as control of action sequences and provides further evidence on the question of global vs. local optimality of human cognition.