Complex decisions made transparent
The third part of our crowd:it reshape improves our agents' decision making. Missed the first two parts of our lockdown reshape? Click here for Part 1 - In 6 steps to evaluation and here for Part 2 - Our new agent behavior.
Agents make decisions
Agents have to make some decisions on the way to their destination, for example, which staircase to take, or which queue to wait in. To be able to model these decisions, there used to be some criteria (heuristics) for the agent to choose from - for example, it could always choose the staircase it is currently closest to, or the one with the fewest people on it.
However, these simple criteria quickly reach their limits when increasingly complex decisions are to be modeled. Imagine you want to model a situation where people only go to certain waiting areas at certain times, for example at a train platform.
Therefore, we have introduced a new selection process that can be used to provide decision support to an agent with unprecedented granularity. Different selection processes can be valid depending on time, several selection modes can be specified within a selection process, and criteria can be defined individually for each selection mode. This makes our new tool extremely powerful, so that almost any situation can be mapped.
Simple and advanced - choose your level
This sounds very complicated and you have concerns to be able to use this tool without further introduction? Don't worry, you can still use the heuristics you are used to.
A set of simulation objects, from which an agent should select one, is represented by so-called sets. We have fundamentally reworked the set dialog. Now you can see at a glance which objects are contained in a set and edit them directly. With our new design, we made sure that existing users can quickly find their way around. If you like the established heuristics, you can continue to use them as usual.
Advanced users can expect a wealth of options with the new selection processes, allowing you to clearly model even the most demanding scenarios.
How does the selection process work?
When an agent selects its next destination, it often has a large number of options to choose from, for example several staircases and escalators. Within a selection process, several selections can be made one after the other in order to remove more and more objects from this set until only one remains at the end. Within a selection, criteria (e.g., distance of the agent to the object, number of agents already approaching this object, available space in the object, an expected waiting time, etc.) can be used to specify very precisely the conditions under which an object is to be selected.
Complex decisions become transparent
The operation of the new selection process is very transparent, as the pathfinding is much more comprehensible than it was with the heuristics. Thus, we provide our users with a powerful tool that allows a variety of new use cases.
For more information on how to use the selections, please refer to our documentation.