A-computational-software-system-to-design-order-_2021_Computers---Operations[1] (1)

1. Introduction to GABAK

• GABAK is an open-source computational software system for warehouse layout design.

• Focuses on optimizing average walking distance for order picking operations.

• Traditional warehouse designs have not changed in 60 years; GABAK aims to innovate this.

• Designed using parameters such as:

  • Rectangular aspect ratio of the floor plan.

  • Number and locations of cross aisles and pick aisles.

  • Location of input/output (depot).

2. Key Features of GABAK

2.1 Algorithms and Components

• The software uses:

  • Data Importing: Allows import of pick list profile data.

  • Layout Creation: Generates warehouse layout as a network.

  • Product Allocation: Allocates storage locations for SKUs.

  • Routing Algorithms: Utilizes exact routing for order pickers.

  • Design Optimization: Employs a meta-heuristic for design improvement.

2.2 Software Accessibility

• GABAK is available under MIT license on GitHub, promoting collaborative improvement and utilization.

3. Literature Review

• Traditional Practices: Most warehouse designs rely on experience, with little innovative software assistance.

• Existing tools focus on specific layouts or limited design functionalities:

  • Roodbergen's model minimizes travel distances but is restricted to traditional layouts.

  • GABAK allows re-designing existing warehouses, providing more flexibility in layouts.

4. Methodology

4.1 General Framework

• A systematic approach identifying superior layout designs:

  • Minimizes travel distance/time following existing operational practices.

  • Involves different floor plan configurations and computational checks against real-world constraints.

4.2 Assumptions

• Key assumptions include:

  • Turnover frequency of SKUs is constant.

  • Only straight-line distances are calculated, ignoring lateral movements within aisles.

  • Each SKU assigned to a unique storage location.

5. Warehouse Design Classes

• GABAK can evaluate 19 distinct design classes:

  • Each class characterizes layout via exterior and interior nodes along with cross aisles.

  • Flexible enough to represent diverse real-world designs.

6. Optimization Model

6.1 Mathematical Formulation

• A modified version of the traditional traveling salesman problem (TSP) to minimize picker distance:

  • Incorporates constraints to enforce unique placements for SKUs.

  • Utilizes both aisle-center and visibility graph methods to calculate distances.

6.2 Evolutionary Strategies Meta-Heuristic

• GABAK employs an Evolutionary Strategy (ES) algorithm for layout optimization:

  • Adaptively adjusts search strategies, improving exploration and exploitation.

  • The default settings are configured to efficiently generate multiple designs.

7. User Interface and Visualization

• GABAK features a graphical user interface (GUI) allowing:

  • Parameter adjustments for warehouse and pick list data.

  • Printable layouts and graphical representations of designs.

  • Animated visualizations to showcase layout evolution during optimization.

8. Validation and Testing

• Validation showed GABAK could produce results comparable to known optimal solutions:

  • Effective performance reported in a controlled environment with varying design tests.

9. Applications and Future Work

• GABAK has applications for researchers, providing a platform for:

  • Assessing the impact of different design strategies.

  • Future enhancements include additional routing algorithms, multiple depot configurations, and augmented reality implementations for testing designs in practical settings.