# Week 7
###### tags: `CS4210-B:Intelligent_Decision_Making_Project`

:::info
- **Date:** June 5, 2020 17:00 (UTC+2)
:::

## Progress update
* More experiments have been running
* First results are processed
* First graphs have been created
* Presentation

## Results
**Capacity**
No influence on any metric

**Poisson**
Significant impact on the solving time , number of branches and solutions (also the reason why we changed the default value to 6).
Interesting is the influence on the apprenticeship time (this is not expected)

**Slack**
Solving time seems somewhat decreasing as expected (more slack allowed easier to solve).

**Speed**
Seems especially influencing with really low values. We need to investigate this, probably because due to integer rounding. (i.e. no speed for a vehicle and thus no solutions)

**Vehicles**
Interesting is the bad runtime with 2 vehicles. Impact on the apprenticeship time is as expected.

**General**
* Random forest apprenticeship does not outperform other means. In terms of branches somewhat similar to random hints.
* Providing the previous route as an initial solution (as helped does) is not always beneficial.
* Interesting is that the constraint solver finds the most solutions but evaluates the least amount of branches.

## Questions

## Next week
The initial planning is as following


For next week we plan to make this concrete with the following steps:
* Write the report
* Run more experiments (also with multilayer perceptron)
* Process the results of the experiments, and create comprehensive graphs