Lecture_06_Decision_Models.Rmd

## Lecture 6: Building Decision Models {#decision_models}
<!-- reference with [Decision Models](#decision_models) -->

Welcome to lecture 6 of **Decision Analysis and Forecasting for Agricultural Development**. We will walk through brief examples and offer the scripts. Feel free to bring up any questions or concerns in the Slack, to [Dr. Cory Whitney](mailto:cory.whitney@uni-bonn.de?subject=[Lecture_3]%20Decision%20Analysis%20Lecture) or to the course tutor.

Decision-makers often wish to have a quantitative basis for their decisions. However,‘hard data’ is often missing or unattainable for many important variables, which can paralyze the decision-making processes or lead decision-makers to conclude that large research efforts are needed before a decision can be made. That is, many variables decision makers must consider cannot be precisely quantified, at least not without unreasonable effort. The major objective of (prescriptive) decision analysis is to support decision-making processes faced with this problem. Following the principles of Decision Analysis can allow us to make forecasts of decision outcomes without precise numbers, as long as probability distributions describing the possible values for all variables can be estimated. 

The `decisionSupport` package implements this as a Monte Carlo simulation, which generates a large number of plausible system outcomes, based on random numbers for each input variable that are drawn from user-specified probability distributions. This approach is useful for determining whether a clearly preferable course of action can be delineated based on the present state of knowledge without the need for further information. If the distribution of predicted system outcomes does not imply a clearly preferable decision option, variables identified as carrying decision-relevant uncertainty can then be targeted by decision-supporting research. This approach is explained in more detail below and in the [model programming seminar](#model_programming).

In this portion of the course we hope to introduce you to the methods and inspire you to follow a few important guidelines in the process of model building. One of the key aspects of model building has to do with making a solid business case for the model before programming. 

<iframe width="560" height="315" src="https://www.youtube.com/embed/bViPWX8_G4A" frameborder="0" allowfullscreen></iframe>

Another important guideline is to start simple then move on to other steps. This ensures that you always have a working model at each of the steps in the process, i.e. starting with a skateboard rather than a car part as in this example from [MetaLab](https://twitter.com/metalab).

!['Skateboard, Bike, Car'](images/Model_stages_car_analogy.png)

<!-- ### Group discussion reading: -->

<!--  - @do_decision_2020 'Decision analysis of agroforestry options reveals adoption risks for resource-poor farmers'. -->

### Bonus: Further reading

- @whitney_probabilistic_2018-1 'Probabilistic decision tools for determining impacts of agricultural development policy on household nutrition'.

 - @ruett_model-based_2020 'Model-based evaluation of management options in ornamental plant nurseries'