---
title: A Mobile Phone Example with four Objectives
categories: moo
toc: false
...

> Back to [Introduction](/MOO/Introduction)

Here's a simple example of a model suitable for optimization using [ClaferMoo](https://github.com/gsdlab/claferMooStandalone).
In order to analyze the model and explore the Pareto front in ClaferMoo Visualizer, click the link below:

```{.clafer .mooviz}
```

## Modeling the Optimization Problem

First, we need to define features with some quality attributes `productivity`, `cost`, `batterylife`, and `security`:

```clafer
abstract Feature
    productivity -> integer
    cost -> integer
    batterylife -> integer
    security -> integer
```

Next, we need to write a model of our product line with some variability regarding the features.
We need to set the values for quality attributes of each feature.

```clafer
abstract MobilePhone
    Connectivity : Feature
        [ batterylife = -12]
        [ productivity = 14]
        [ security = 43]
        [ cost = 101]        
        xor Bluetooth
            Bluetooth20EDR : Feature
                [ batterylife = -4]
                [ productivity = 1]
                [ security = -15]
                [ cost = 1]
            Bluetooth21EDR : Feature
                [ batterylife = -2]
                [ productivity = 4]
                [ security = -10]
                [ cost = 1]
            Bluetooth40 : Feature
                [ batterylife = -1]
                [ productivity = 16]
                [ security = -2]
                [ cost = 2]
        GSM : Feature
            [ batterylife = -2]
            [ productivity = 2]
            [ security = -10]
            [ cost = 1]
        LTE : Feature ?
            [ batterylife = -1]
            [ productivity = 16]
            [ security = -3]
            [ cost = 3]
            [ LiBattery1400 ]
        WiFi: Feature ?
            [ batterylife = -10]
            [ productivity = 20]
            [ security = -15]
            [ cost = 20]
        USB : Feature ?
            [ batterylife = 10]
            [ productivity = 20]
            [ security = 0]
            [ cost = 3]
    xor Battery
        LiBattery1150 : Feature
            [ batterylife = 60]
            [ productivity = 0]
            [ security = 0]
            [ cost = 10]
        LiBattery1400 : Feature
            [ batterylife = 70]
            [ productivity = 0]
            [ security = 0]
            [ cost = 15]
   
    total_productivity -> integer = sum Feature.productivity
    total_batterylife -> integer = sum Feature.batterylife 
    total_security -> integer = sum Feature.security
    total_cost -> integer = sum Feature.cost
```

We could define a specialization and extension like this (currently not supported)

```clafer
abstract HighEndPhone : MobilePhone
   [ WiFi ]
   [ Bluetooth40 ]
   QuadCore : Feature ?
       [ batterylife = -10]
       [ productivity = 30]
       [ security = 0 ]
       [ cost = 25 ] 
```

Next, we define our product `OptimalPhone`. 

```clafer
OptimalPhone : MobilePhone
```

The product is underspecified allowing for all possible valid combinations of features.

Finally, we define the following four optimization goals.

"Maximize total battery life":

```clafer
<< maximize OptimalPhone.total_batterylife >>
```

"Maximize total productivity":

```clafer
<< maximize OptimalPhone.total_productivity >>
```

"Maximize total security":

```clafer
<< maximize OptimalPhone.total_security >>

```

"Minimize total cost":

```clafer
<< minimize OptimalPhone.total_cost >>
```

In order to analyze the model and explore the Pareto front in ClaferMoo Visualizer, click the link below:

```{.clafer .mooviz}
```

## Comparing existing products

Engineers may want to compare the existing product variants (e.g., currently in production) with the optimal variants available in the new version of the product line (e.g., after the feature `LTE` has been added).

For example, consider an existing product `MyOldPhone` which has an older version of Bluetooth, `20EDR`, and no `WiFi`, `LTE`, nor `USB`.
Here is the  configuration of that product in Clafer notation:

```
MyOldPhone : MobilePhone 
    [ Bluetooth20EDR ]
    [ no LTE ]
    [ no WiFi ]
    [ no USB ]
    [ LiBattery1150 ]
```

In order to load that configuration into the visualizer

1. Download the [instance data file](/MOO/MobilePhone.cfr.1.data) corresponding to the above instance specification,
2. In Visualizer, open that file in the `Inputs` view and click on the button `Add Instances`.
* The visualizer will load this configuration and show it alongside the optimal products as a square.

Now, you may proceed with some exploration. For example, 

1. Select the existing variant `17` and the closest optimal variant `2`.
* The visualizer will show the differences in the view `Comparison`
2. Observe that changing the feature `Bluetooth20EDR` to `Bluetooth21EDR` would make the product optimal
3. Set the upper bound for `cost` to `120` in the view `Objectives`
* The cost `120` is the largest acceptable in the market segment
* The visualizer only shows variants costing less or equal to `120`
4. Select the best variant `16` for comparison
* Observe that by increasing cost by just `7`, we can obtain a far superior variant which includes the latest `Bluetooth40`, `LTE`, and `USB` but no `Wifi`.