Collaborative variable neighborhood search for multi-objective distributed scheduling in two-stage hybrid flow shop...

Area of Science:

• Operations Research
• Industrial Engineering
• Manufacturing Systems

Background:

• Distributed scheduling in hybrid flow shops is an under-researched area.
• The distributed two-stage hybrid flow shop scheduling problem (DTHFSP) with sequence-dependent setup times presents significant complexity.
• Existing scheduling methods often do not address the distributed nature and sequence-dependent setups simultaneously.

Purpose of the Study:

• To develop an effective scheduling algorithm for the DTHFSP.
• To simultaneously minimize two critical performance metrics: total tardiness and makespan.
• To introduce a novel solution representation for the DTHFSP.

Main Methods:

• A Collaborative Variable Neighborhood Search (CVNS) algorithm is proposed.
• The DTHFSP is simplified by integrating factory and machine assignment.
• CVNS utilizes two cooperating Variable Neighborhood Search (VNS) algorithms with eight neighborhood structures and two global search operators.
• A dynamic archive mechanism replaces solutions with the farthest member.

Main Results:

• The proposed CVNS algorithm demonstrates significant advantages in solving the DTHFSP.
• Computational experiments validate the effectiveness of CVNS in minimizing both total tardiness and makespan.
• The novel solution representation facilitates efficient search and problem simplification.

Conclusions:

• CVNS is a highly effective approach for the complex DTHFSP.
• The collaborative nature of VNS algorithms within CVNS enhances search efficiency.
• The study provides a valuable contribution to distributed scheduling research in hybrid flow shops.