With the ever-increasing product variety faced by the manufacturing industry, investment efficiency can only be maintained by the application of multi-product assembly systems. In such systems, the product design, process planning, and production planning problems related to different products are strongly interconnected. Despite this, those interdependent decisions are typically made by different divisions of the company, by adopting a decomposed planning approach, which can easily result in excess production costs. In order to overcome this challenge, this paper proposes an integrated approach to solving the above problems, focusing on the decisions crucial for achieving the required tolerances in high-precision assembled products. The joint optimization problems related to product tolerance design and assembly resource configuration are first formulated as a mixed-integer linear program (MILP). Then, a large neighborhood search (LNS) algorithm, which combines classical mathematical programming and meta-heuristic techniques, is introduced to solve large instances of the problem. The efficiency of the method is demonstrated through an industrial case study, both in terms of computational efficiency and industrial effectiveness.

Introduction and motivation

In response to diversifying consumer preferences, many companies from the automotive, electronics, and consumer goods industries are forced to increase product variety [1–۳]. The situation is often complicated further by the changes of the conventional manufacturer-supplier relationships, e.g., in the automotive industry, where a single supplier now serves many manufacturers. Therefore, the supplier must increase its product variety, and the demand for multi-variety production grows. As a consequence, requirements of new products often cannot be satisfied by existing manufacturing and assembly lines, and therefore, investment into new equipment is inevitable. There are also attempts to lift manufacturing constraints by introducing general purpose equipment, but excessive generalization or flexibility of equipment can also lead to low production rate and low return on investments [4]. In the conventional product development process, different phases of the process focus on different issues to be resolved: first of all, product design has to meet customer specifications by selecting appropriate design alternatives. When a product design is available, process planning is responsible for realizing the design by defining the assembly resource configurations. In the operation stage, production planning assigns products to resources over time to satisfy demand in the most efficient way.