Firefly Semantics performs Service Parts Profit Optimization in two stages, described below.

The **FS Alpha** Product handles **Stage I** only.

The primary purpose of this product is to be a learning tool, all though local optimization does provide business insight.

**FS Beta**, which will be released soon, handles **Stage II**, and **Stage I**. It shows planners the most valuable planning scenarios to focus on.

**FS Gamma** will perform Budget Constraint Optimization Analysis when the possibility of affording optimal stocking levels for all parts is not possible.

# { Stage I } — Local Optimization

The first stage is Local Optimization where we focus on determining the Optimal Stocking Level for the following inputs:

- Annual Demand level
- Purchase Cost
- Sales Price
- Cost of Capital
- Lead Time Days

## Step 1

Construct the Demand Distribution modeling demand probability over the lead time days for the part.

## Step 2

For each stocking level calculate:

## Step 3

Check whether the Marginal Profit is greater than the Marginal Inventory Cost

If it is increment the stocking level and repeat the steps.

If it is not the previous stocking level was the Optimal Stocking Level.

# { Stage II } — Global Optimization

In Stage I we considered a fixed annual demand level.

In Stage II we will consider annual demand levels ranging from 1 to 24

## Step 1

Develop the Prior Probability Distribution.

## Step 2

Develop the Posterior Probability Distribution

## Step 3

In this step we score all possible demand scenarios ranging from 1 to 24.

For each demand level the optimal stocking level is calculated.

We then calculate the net profit at the optimal stocking level for each demand scenario.

We then score the net profit calculation by multiplying it by the probability of being in that scenario.

We use both probability distributions for this.

So for example for a mean annual demand level of `10`

units, we have an expected net profit of $44,000,000.

The probability of this scenario occurring is `0.14`

.

The score is `0.14*44000000`

or **6,160,000**.

For a mean annual demand level of 23 units, we may have an expected net profit of $72,000,000.

The probability of this scenario occurring is 0.001.

The score is **72,000**.

Which scenario should the planner focus on?