Optimization of the Design of Modular Production Systems

Introduction

The customer's increasing desire for more individuality in recent years presents companies with new challenges, which are further intensified by shorter innovation cycles and, as a result, often shorter product cycles, as well as external influences from the market, the state, etc.

Consequently, an essential requirement for modern production systems, especially in final assembly, is to be able to react flexibly to changing conditions while maintaining the economic efficiency of production for the sometimes large number of products or product variants.

This paper presents the concept of simulation-based optimization of the system design of a modular assembly system, which is demonstrated and validated in the next section by means of an academic case study.

Optimization model

The concept for simulation-based optimization of the system design of modular assembly and production systems (cf. Figure) has been designed as generically as possible in order to be usable in many applications.


Basis concept for simulation-based optimization of system design for modular product systems
Basis concept for simulation-based optimization of system design for modular product systems

The basis was a suitable simulation model with the corresponding basic layout of the production, which had suitable interfaces.

In this paper, for the simulation, an agent-based approach implemented in AnyLogic was used, and as the optimization tool, HeuristicLab was chosen because it supported many optimization algorithms, such as the classical genetic algorithm (GA) and the non-dominated sorting genetic algorithm II (NSGA-II) algorithm.


Schematic flow of the optimization
Schematic flow of the optimization

Results

This paper presented a concept for simulation-based optimization of the system design of modular production systems using a classical GA and the NSGA-II algorithm, which supported a good activity assignment to the individual manufacturing cells.

The specifics of modular production systems were first discussed before describing the basic features of simulation-based optimization using genetic algorithms and the multi-objective NSGA-II algorithm. A simple academic case study was used to demonstrate the concept and its prototypical implementation.

Related posts