A Hybrid SLP-MILP Approach for Facility Layout Optimization in Specialized Manufacturing Environments

Date of Award

6-2026

Degree Name

Master of Science

Department

Industrial and Entrepreneurial Engineering and Engineering Management

First Advisor

Ilgin Acar, Ph.D.

Second Advisor

Lee Wells, Ph.D.

Third Advisor

Sang Kang, Ph.D.

Access Setting

Masters Thesis-Abstract Only

Restricted to Campus until

6-1-2036

Abstract

Facility Layout Planning (FLP) is a critical determinant of production efficiency and material handling cost in manufacturing systems. Existing solution approaches fall into three paradigms: graph-based methods that capture adjacency relationships without guaranteeing geometric feasibility, guillotine-cut partitioning methods that ensure spatial feasibility without incorporating relational priorities, and MILP-based methods that optimize within a feasible region without structured domain-knowledge initialization. Each paradigm addresses only a partial aspect of the problem, to the best of author’s knowledge, no prior formulation integrates all three within a unified approach.

This thesis proposes a Hybrid SLP-MILP methodology integrating all three paradigms into a sequential three-phase pipeline. The first phase computes V-scores from Systematic Layout Planning (SLP) closeness ratings to determine machine placement priorities and zone assignments. The second phase applies a two-dimensional guillotine-cut partitioning algorithm to generate a geometrically feasible warm-start configuration. The third phase applies a Mixed- Integer Linear Programming (MILP) model enforcing ten constraint groups, including disjunctive non-overlap, production-sequence ordering, Cleanroom touch requirements, and adjacency penalties, within a six-term composite objective function. The CBC solver certifies optimality within a 1% gap tolerance for the 458-variable, 696-constraint formulation, meaning the solution is within 1% of the best-known optimum. Applied to the 11-machine PCB manufacturing facility at the WMU CEAS Annex, the Hybrid SLP-MILP achieves 10/10 Grade- A adjacency satisfaction, a valid SMT production sequence, confirmed Cleanroom touch for both constrained machines, bounding box compactness 70.2 ft and composite objective Z = 561.996 in 160.49 seconds. Among all four evaluated configurations, it is the only one that is simultaneously operationally feasible, constraint-complete, and near-optimal.

Preliminary evidence of generalizability was obtained by applying the fourconfiguration comparison to three external instances: the NUG12 benchmark (Nugent et al., 1968), the SCR12 benchmark (Scriabin and Vergin, 1975), and a synthetic automotive manufacturing instance (SYN-AUTO12) developed as part of this study. Across all three instances, the Hybrid SLP-MILP consistently outperforms heuristic baselines in material handling cost and adjacency satisfaction. On instances with domain-specific hard constraints, it is the only configuration that produces a globally optimal and operationally feasible solution.

This document is currently not available here.

Share

COinS