Skip to main content

Hi,

I’m wondering if the machine learning part of the SF can be used - and if so, how - for the following scenario.

I’m thinking of creating something of a Planning tool. I have a table with people with (several) skills (truck driver, crane operator,..) , a table with tools (e.g truck, broom, crane, etc) and a table with jobs that need to be done on a certain day/time. 

Say, a truck needs to be loaded by 3 people and driven by 1. A ship needs to be unloaded by 3; a crane operator, coordinator, lorry driver. The dock needs to be cleaned by 5 people. We have the tools needed available listed. 

I suspect that the machine learning part can - in time - do some math and suggest how the right people can do all these jobs in the most efficient way. 

Is this correct? And what do I need to keep in mind for creating something like this? 

Curious if this might be something to look into! 

Thanks! 

At the time of writing, this is not something our automated machine learning software can do.

Currently, the supported problem types are regression and classification. These correspond roughly to fitting lines through data and assigning classes to them. Solutions to these problems consist of a model that generates a single number per data point (a continuous number corresponding to the regression line, or an integer corresponding the predicted class). A planning on the other hand is a solution with more structure. We do not at the moment have any way of generating models that generate such solutions.

I also think that planning is unlikely to be added in the near future.

Creating specific planning tools is very much something we do. But in general, planning is an unsolved problem, meaning that there is no general recipe for finding (models that propose) good solutions. That means that it doesn’t really fit with our automated machine learning software.

Terms & ConditionsCookie settings