Various computer programs facilitate the work of urban planners. Esri CityEngine, for instance, enables to analyze and compare building proposals from every angle and creates awesome 3D visualizations. Autodesk offers even more than one software solution. Autodesk InfraWorks, for example, supports transportation or site planning and even water analysis. Those interested in statistical data analysis are better advised with SAS/SPSS/STATA. It is, of course, also possible to decide for ArcGIS, Adobe Design Creative Suite, SketchUp, Maya, 3D Studio Max and so on and so forth. Is there a planning challenge which can’t be tackled with a certain software?
In some cases, surveying engineers are already replaced by drones
If no, what kind of software is recommended to replace a costly urban planner? You don’t take it very seriously? In this case, it’s worth remembering technical drawers with their outdated drawing boards – which may nowadays only be useful for museums. Another topical example are surveying engineers. More and more, they get replaced by drones. They are excellent for mapping roads, highways, and railway infrastructures – drones can in many ways do the job faster, cheaper and safer than surveying engineers.
However, urban planners will assuredly not be redundant overnight. Instead, it is much more a question of how many and which working packages will be replaced by algorithms and robots in the future. The imitation of human behavior with a much better performance through artificial intelligence has been already showcased in a number of certain tasks.
The next step is to teach machines to do the same thing
It’s official. When it comes to solving complex problems, a smart human is still able to find better solutions than a machine. What’s the magic ingredient? It turns out that the missing talent that could make planning algorithms more efficient is linear temporal logic says a group of MIT researchers.
The group presented its findings at the annual conference of the Association for the Advancement of Artificial Intelligence, and the work will be ongoing. Now that we know why people are better at solving complex problems, the next step is to teach machines to do the same thing.
The researchers have begun their work with automated planning algorithms that can help with problems such as efficient flight route planning, saving fuel, reducing fight time, and maximizing the number of passengers on each plane. The principles used to do this could then be applied to other forms of complex problem-solving.
How the New Planning Algorithms Fared With Input From MIT Students
To test the performance of International Planning Competition (IPC) award-winning planning algorithms, the researchers pitted them against MIT students with high-level problem-solving skills. Three classes of problem-solving were tested: the ability to solve a problem while satisfying a rigid set of constraining factors, numerical problems with parameters that include a degree of flexibility, and temporal problems that added a time-based element to the exercise.
Both the algorithms and the MIT students were given half an hour to produce a solution. When evaluating the student’s solutions, the researchers interviewed them, observing that linear temporal logic was widely used to add constraints that had not been specified when the problem was initially defined.
By adding these constraints to the code, the plans produced by algorithms were substantially improved, with researchers identifying a 10 to 16 percent improvement in the planning algorithm’s ability to solve the various types of problems, and the researchers noted that the solutions were much closer to those that had been calculated by human beings.
“In the lab, in other investigations, we’ve seen that for things like planning and scheduling and optimization, there’s usually a small set of people who are truly outstanding at it,” says Julie Shah, an assistant professor of aeronautics and astronautics at MIT. “Can we take the insights and the high-level strategies from the few people who are truly excellent at it and allow a machine to make use of that to be better at problem-solving than the vast majority of the population?”
Talk to the Machine: Turning Strategies into Code
Since situations requiring help from planning algorithms are affected by a multitude of factors, high-level strategies used in problem solving would vary. But with natural language processing techniques having reached an advanced level, the MIT researchers believe that a free-form description of a problem-solving strategy could be transformed into the kind of linear temporal logic that makes sense to an algorithm. The hope-for result? Improved planning algorithms that will make use of the best both human and machine logic can contribute.