“Search algorithms are essentially game solvers. They compute data and solve a problem space. These algorithms look at different possibilities and constraints and use various methods to find optimal paths and ultimately solve problems.”
Lists --A new approach to scheduling
A common theme regardless of the technique is the concept of using lists. In order to use search in scheduling, a new framework for assigning start and end dates to tasks is needed. The current methods for scheduling are based on ideas from the 1960s, and as such need to be rethought.
By examining different possibilities in different ways, search algorithms use techniques to compile lists. Lists are sequenced and examined until a decision is reached. The final list, or done list, is a solution (or schedule) that is produced with the artificial intelligence. For construction schedules, four lists can be used to describe tasks, similar to the kan-ban approach famously developed by Toyota.
The first type of list is the to-do list. With a to-do list, you're planning all your tasks and constraints with their associated durations, relationships, and resources. These lists define the rules governing the construction plan.
The next type of list is the can-do list. Here, the algorithm resolves constraints. For example, one task must start before another task the other task cannot start before its required resources are available. Altogether, such a resolution accumulates into a can-do list.
The third type is a doing list. A doing list is your future-state generation, which is actually the journey one travels towards the end goal or your done list. The start and end time of this task is assigned. Then resources are assigned from the resource pool. After that, they move to the next task in the can-do list. Tasks are selected to move from a can-do list to the doing list, while artificial intelligence considers the different sequence possibilities.
The last list is the done list. The schedule is complete when all the tasks from the to-do list are on the done list.
Search algorithms not only use different techniques to explore the various possibilities and resolve a problem, they also use a list approach to get there. The list approach is sequential. It involves different constraints and dependencies.
Algorithms and Construction Scheduling
When applied properly, search algorithms and their application to ranked lists can be a boon to the construction scheduler. They make a major difference in reducing the construction schedule with respect to multiple alternatives and many constraints.
The resource pool may include everything from workers to cranes, machines, and other tools and equipment used for the job. However, ultimately the done list recommends available labor teams that have the availability to perform tasks according to the best schedule.
In essence, the use of search algorithms is about to revolutionize construction scheduling via software solutions that provide an entirely new approach, with attractive results. Smart construction schedulers who embrace and engage these tools stand to have a distinct advantage over traditional and proverbial construction scheduling tools, technologies, and solutions.
When a 2017 article by McKinsey, “Construction: The Next Great Tech Transformation," discussed a new model for construction that incorporates technology in all its facets, they described it as presenting "an exciting opportunity and a big win for everyone in the industry. We can remove time and costs to the great benefit of consumers, create a new generation of skilled manufacturing jobs, and modernize an industry that is vitally important to our social fabric and to the planet."
The use of search algorithms within artificial intelligence to upgrade construction scheduling is doing just that: removing time and costs while modernizing the construction industry.