Block Diagram Rule Base 1

• Inputs: Location, Market Value (Yellow Blocks)

• Process: Mamdani FIS (Fuzzification, Rule Evaluation, Defuzzification)

• Output: House Suitability (Blue Block)

Rule Viewer

• Inputs: Location = 5, Market Value = 525

• Output: House Suitability = 7

• Visualization: Degrees of membership shown in yellow; output shown in blue.

Surface Viewer

• Axes: X and Y (input variables), Z (output variable)

• Visualization: 3D plot to identify patterns and relationships

Mortgage 2

Fuzzy Logic Designer

• Inputs: Asset (0-1000), Income

• Output: Applicant Eligibility

• Methods: Min (AND), Max (OR), Centroid (Defuzzification)

• Rules: 12

Rule Viewer

• Inputs: Asset = 500, Income = 55

• Output: Applicant Eligibility = 8.37

• Visualization: Degree of membership in yellow; output calculation.

Surface Viewer

• Axes: X (Asset), Y (Income), Z (Applicant Eligibility)

• Visualization: 3D plot showing impact of asset and income on eligibility

Mortgage 3

Fuzzy Logic System

• Inputs: Income, Interest, Applicant, House (Yellow Blocks)

• Output: Credit (Blue Block)

• Methods: Min (AND), Max (OR), Centroid (Defuzzification)

• Rules: 15

Rule Viewer

• Inputs: Income = 55, Interest = 5, Applicant = 5, House = 5

• Output: Credit Score = 250

• Visualization: Degree of rule activation in yellow; final defuzzified output in blue.

Surface Viewer

• Grid: X Mesh 15, Y Mesh 15

• Rules: 15

• Visualization: 3D plot showing loan risk based on input variables

Retail stores often face challenges with inventory management, leading to stockouts or overstocking. Stockouts cause missed sales and unhappy customers, while overstocking ties up capital and increases storage costs.

Key Variables:

1. Inventory Levels: Current stock of each product.

2. Demand: Projected sales over a specific period.

3. Lead Time: Time taken to replenish stock.

4. Storage Capacity: Total space available for storing products.

5. Cost: Expenses related to holding inventory.

Fuzzy Sets:

• Inventory Levels: Low, Medium, High

• Demand: Low, Medium, High

• Lead Time: Short, Medium, Long

• Storage Capacity: Limited, Medium, Abundant

• Cost: Low, Medium, High

Decision Rules:

1. Low Inventory & High Demand: Order more stock.

2. High Inventory & Low Demand: Reduce inventory.

3. Short Lead Time & Abundant Storage: Hold more inventory.

4. High Cost & Limited Storage: Reduce inventory.

Implementation Steps:

1. Data Collection: Gather accurate data on inventory levels, demand forecasts, lead times, storage capacities, and holding costs.

2. Data Analysis: Use fuzzy logic to handle uncertainties and define fuzzy sets.

3. Decision Rules: Implement rules in the DSS to generate recommendations.

4. User Interface: Develop an intuitive interface for store managers to input data and receive actionable recommendations.

Benefits:

• Reduces stockouts and overstocking.

• Improves customer satisfaction.

• Optimizes storage costs.

• Handles uncertainties in inventory management effectively.

By leveraging a DSS with fuzzy logic, retail stores can make informed and efficient inventory decisions, enhancing overall performance and profitability.

This guide explains how to build a fuzzy expert system in MATLAB for making inventory management decisions. The system uses fuzzy logic to evaluate various input parameters and make decisions about inventory levels.

Step-by-Step Breakdown:

1. Launch the Fuzzy Logic Designer:

   • Use the fuzzyLogicDesigner command to open the Fuzzy Logic Designer app in MATLAB. This graphical interface helps create and manage fuzzy inference systems (FIS).

2. Create a New FIS:

   • Define a new FIS named 'Inventory Management' using fis = newfis('Inventory Management');.

3. Define Input Variables:

   • Five key input variables are defined, each with a specific range:

     1. Inventory Levels: 0 to 100

     2. Demand: 0 to 200

     3. Lead Time: 0 to 30

     4. Storage Capacity: 0 to 100

     5. Cost: 0 to 30

4. Add Input Variables to FIS:

   • Use the addvar function to add each input variable to the FIS with its respective range.

5. Define Membership Functions (MFs):

   • Each input variable is categorized into three linguistic terms using triangular membership functions (trimf):

     • Inventory Levels: 'Low', 'Medium', 'High'

     • Demand: 'Low', 'Medium', 'High'

     • Lead Time: 'Short', 'Medium', 'Long'

     • Storage Capacity: 'Limited', 'Adequate', 'Abundant'

     • Cost: 'Low', 'Medium', 'High'

6. Define Output Variable:

   • The output variable 'Inventory Management Decision' ranges from 0 to 100 and has three MFs: 'Order More', 'Hold Inventory', 'Reduce Inventory'.

7. Set Fuzzy Rules:

   • Define rules that govern the decision-making process using a matrix where each row represents a rule. For example:

     • 1 1 1 1 1 1 1 means if 'Inventory Levels' is 'Low' and 'Demand' is 'High', then 'Order More'.

   • Use the addrule function to add these rules to the FIS.

8. Save the FIS:

   • Save the system to a file using savefis(fis, 'inventory_management.fis'); for future use.

9. Create a MATLAB App for Evaluation:

   • Design a GUI with input fields for each variable, and an 'Evaluate' button.

   • Implement a callback function evaluate_callback that retrieves input values, evaluates the FIS using evalfis, and displays the decision.

10. Run the App:

    • Use the uiwait function to keep the app running and wait for user interaction.

Test Cases:

1. Test Case 1:

   • Inputs: Inventory Levels = 15, Demand = 180, Lead Time = 3, Storage Capacity = 90, Cost = 8

   • Expected Decision: 'Order More'

   • Scenario: Low inventory, high demand, short lead time, abundant storage, and low cost.

2. Test Case 2:

   • Inputs: Inventory Levels = 85, Demand = 40, Lead Time = 18, Storage Capacity = 60, Cost = 25

   • Expected Decision: 'Reduce Inventory'

   • Scenario: High inventory, low demand, long lead time, limited storage, and high cost.

• Show the new Fuzzy Inference System (FIS) for edge detection.