Multi-objective Evolutionary Optimisation for Product Design and Manufacturing

As the name suggests, multi-objective optimisation involves optimising a number of objectives simultaneously. The problem becomes challenging when the objectives are of conflicting characteristics to each other, that is, the optimal solution of an objective function is different from that of the other. In the course of solving such problems, with or without the presence of constraints, these problems give rise to a set of trade-off optimal solutions, popularly known as Pareto-optimal solutions. Because of the multiplicity in solutions, these problems were proposed to be solved suitably using evolutionary algorithms using a population approach in its search procedure. Starting with parameterised procedures in early 90s, the so-called evolutionary multi-objective optimisation (EMO) algorithms is now an established field of research and application with many dedicated texts and edited books, commercial softwares and numerous freely downloadable codes, a biannual conference series running successfully since 2001, special sessions and workshops held at all major evolutionary computing conferences, and full-time researchers from universities and industries from all around the globe. In this chapter, we provide a brief introduction to its operating principles and outline the current research and application studies of evolutionary multi-objective optmisation (EMO).

Research regarding supply chain optimisation has been performed for a long time. However, it is only in the last decade that the research community has started to investigate multi-objective optimisation for supply chains. Supply chains are in general complex networks composed of autonomous entities whereby multiple performance measures in different levels, which in most cases are in conflict with each other, have to be taken into account. In this chapter, we present a comprehensive literature review of existing multi-objective optimisation applications, both analytical-based and simulation-based, in supply chain management publications. Later on in the chapter, we identify the needs of an integration of multi-objective optimisation and system dynamics models, and present a case study on how such kind of integration can be applied for the investigation of bullwhip effects in a supply chain.

During the last couple of decades the possibility of modelling multi-physics phenomena has increased dramatically, thus making simulation of very complex manufacturing processes possible and in some fields even an everyday event. A consequence of this has been improved products with respect to properties, weight/stiffness ratio and cost. However this development has mostly been based on “manual iterations” carried out by the user of the relevant simulation software rather than being based on a systematic search for optimal solutions. This is, however, about to change because of the very tough competition between manufacturers of products in combination with the possibility of doing these highly complex simulations. Thus, there is a crucial need for combining advanced simulation tools for manufacturing processes with systematic optimisation algorithms which are capable of searching for single or multiple optimal solutions. Nevertheless, despite this crucial need, it is interesting to notice the very limited number of contributions in this field and consequently this makes us wonder about the underlying reasons for it. The understanding of the physical phenomena behind the processes, the current numerical simulation tools and the optimisation capabilities which all mainly are driven by the industrial or academic demands as well as computational power and availability of both the simulation and the multi-objective optimisation oriented software on the market are the main concerns to look for. These limitations eventually determine what is in fact possible today and hence define what the “state-of-the-art” is. So, seen from that perspective the very definition of the state-of-the-art itself in the field of optimisation of manufacturing processes constitutes an important discussion. Moreover, in the major research fields of manufacturing process simulation and multi-objective optimisation there are still many issues to be reserved.

Product family design involves the development of multiple products that share common components, modules and subsystems, yet target different market segments and groups of customers. The key to a successful product family is the product platform—the common components, modules and subsystems—around which the family is derived. The fundamental challenge when designing a family of products is resolving the inherent trade-off between commonality and performance. If there is too much commonality, then individual products may not meet their performance targets; however, too little sharing restricts the economies of scale that can be achieved during manufacturing and production. Multi-objective evolutionary optimisation algorithms have been used extensively to address this trade-off and determine which variables should be common (i.e., part of the platform) and which should be unique in a product family. In this chapter, we present a novel approach based on many-objective evolutionary optimisation and visual analytics to resolve trade-offs between commonality and many performance objectives. We provide a detailed example involving a family of aircraft that demonstrates the challenges of solving a 10-objective trade-off between commonality and the nine performance objectives in the family. Future research directions involving the use of multi-objective optimisation and visual analytics for product family design are also discussed.

Functional commonalities across product families have been considered by a large body of product family design community but this concept is not widely used in design. For a designer, a functional family refers to a set of designs evaluated based on the same set of qualities; the embodiments and the design spaces may differ, but the semantics of what is being measured (e.g., strength of a spring) remain the same. Based on this functional behaviour we introduce a product family hierarchy, where the designs can be classified into phenomenological design family, functional part family and embodiment part family. And then, we consider the set of possible performances of interest to the user at the embodiment level, and use multi-objective optimisation to identify the non-dominated solutions or the Pareto-front. The designs lying along this front are mapped to the design space, which is usually far higher in dimensionality, and then clustered in an unsupervised manner to obtain candidate product groupings which the designer may inspect to arrive at portfolio decisions. We highlight and discuss two recently suggested techniques for this purpose. First, with help of dimensionality reduction techniques, we show how these clusters in low-dimensional manifolds embedded in the high-dimensional design space. We demonstrate this process on three different designs (water faucets, compression springs and electric motors), involving both continuous and discrete design variables. Second, with the help of a data analysis of Pareto-optimal solutions, we decipher common design principles that constitute the product portfolio solutions. We demonstrate this so-called ‘innovization’ principles on a spring design problem. The use of multi-objective optimisation (evolutionary and otherwise) is the key feature of both approaches. The approaches are promising and further research should pave their ways to better design and manufacturing activities.

Product family design is a well recognised method to address the demands of mass customisation. A potential drawback of product families is that the performance of individual members are reduced because of the constraints added by the common platform, i.e., parts and components need to be shared by other family members. This chapter presents a framework where the product family design problem is stated as a multi-objective optimisation problem and where multi-objective evolutionary algorithms are applied to solve the problem. The outcome is a Pareto-optimal front that visualises the trade-off between the degree of commonality (e.g., number of shared components) and performance of individual family members. The design application is a family of industrial robots. An industrial robot is a mechatronic system that comprises a mechanical structure (i.e., a series of mechanical links), drive-train components (including motors and gears), electrical power units and control software for motion planning and control.

In this chapter, we methodologically describe a multi-objective problem solving approach, concurrently minimising two conflicting goals—average surface roughness—Ra and build time—T, for object manufacturing in Fused Deposition Method (FDM) process by usage of evolutionary algorithms. Popularly used multi-objective genetic algorithm (NSGA-II) and recently proposed multi-objective particle swarm optimisation (MOPSO) algorithms are employed for the optimisation purposes. Statistically significant performance measures are employed to compare the two algorithms and approximate the Pareto-optimal fronts. To refine the solutions obtained by the evolutionary optimisers, an effective mutation-driven hill-climbing local search is proposed. Three new proposals and several suggestions pertaining to the issue of decision making in the presence of multiple optimal solutions are made. The overall procedure is integrated into an engine called MORPE—multi-objective rapid prototyping engine. Sample objects are considered and several case studies are performed to demonstrate the working of MORPE. Finally, a careful investigation of the optimal build orientations for several considered objects is done or selected basis and a trend is discovered, which can be considered highly useful for various practical rapid prototyping (RP) applications.

In this study, an ant colony optimisation (ACO)-based setup planning system focusing on an integrated procedure for automatic setup planning for machining cast parts is presented. It considers the selection of available machine tools, tolerance analysis and cost modelling simultaneously for achieving an optimal setup planning result. A tolerance cost factor is introduced when machining error stack-up occurs. The setup planning process can be divided into three stages: preliminary setup planning, tolerance planning and optimal setup planning. During the preliminary setup planning stage, design information is extracted from CAD models and each machining feature is assigned certain machine resource based on its tool access directions (TAD) and the tool orientation space of the available machine resource. During the tolerance planning stage, machining features are grouped into setups based on machine tools assigned and their TADs, and the machining datum for each setup is determined. The setups are next sequenced. Then the blueprint tolerances of the machining features are checked based on their ideal setup datum, and a tolerance cost factor is generated accordingly. During the optimal setup planning stage, the manufacturing cost of each setup plan is evaluated based on the cost model, in which, multiple objectives (setup change cost, machine tool cost, cutter change cost, etc.) that are possibly in conflict with each other are combined through the use of a weight vector and an aggregation function. The setup plan which incurs the least cost is taken as the final result. The feasibility of using the ACO algorithm is studied to address the NP-complete setup planning problem. A case study is carried out to illustrate the proposed approach. This approach can optimise product design and its manufacturing processes simultaneously to meet cost, time and performance objectives, achieving product quality and user satisfaction.

Traditionally, scheduling problems usually deal with the objectives related to production efficiency (e.g., the make-span, the total completion time, the maximum lateness and the number of tardy jobs). However, sustainable manufacturing should minimise the energy consumption during production process. Energy consumption not only constitutes a major portion of total production cost but also results in significant environmental effects. In this chapter, we discuss a multi-objective scheduling problem in a hybrid flow shop. Two objectives considered in the proposed model are to minimise make-span and energy consumption. These two objectives are often in conflict with each other. A Preference Vector Ant Colony System (PVACS) is developed to search for a set of Pareto-optimal solutions using meta-heuristics for multi-objective optimisation. PVACS allows the search in the solution space to focus on the specific areas which are of particular interest to decision-makers, instead of searching for the entire Pareto frontier. This is achieved by maintaining a separate pheromone matrix for each objective, respectively and assigning each ant a preference vector that represents the preference between the two objectives of the decision-makers. The performance of PVACS was compared to two well-known multi-objective genetic algorithms: SPEA2 and NSGA-II. The experimental results show that PVACS outperforms the other two algorithms.

Traditionally, process planning and scheduling were performed sequentially, where scheduling was executed after process plans had been generated. Considering the fact that the two functions are usually complementary, it is necessary to integrate them more tightly so that the performance of a manufacturing system can be improved greatly. In this chapter, a multi-agent-based framework has been developed to facilitate the integration of the two functions. In the framework, the two functions are carried out simultaneously, and an optimisation agent based on evolutionary algorithms is used to manage the interactions and communications between agents to enable proper decisions to be made. To verify the feasibility and performance of the proposed approach, experimental studies conducted to compare this approach and some previous works are presented. The experimental results show the proposed approach has achieved significant improvement.

Autonomous Distributed Manufacturing Systems (ADMS) have been proposed to realise flexible control structures of manufacturing systems. In the previous researches, a real-time scheduling method based on utility values has been proposed and applied to the ADMS. Multi-agent reinforcement learning is newly proposed and implemented to the job agents and resource agents, in order to improve their coordination processes. The status, the action and the reward are defined for the individual job agents and the resource agents to evaluate the suitable utility values based on the status of the ADMS. Some case studies of the real-time scheduling have been carried out to verify the effectiveness of the proposed methods.

Manufacturing scheduling is a well-known complex optimisation problem. A flexible manufacturing system on one side eases the manufacturing processes but on the other hand it increases the complexity in the decision making processes. This complexity further enhances when disruption in the manufacturing processes occurs or when arrival of new orders is considered. This requires rescheduling of the whole operation, which is a complex decision making process. Realising this complexity and taking into account the contradictory objective of making a trade-off between costs and time, this research aims to generate an effective manufacturing schedule. The existing approach of rescheduling sometimes generates entirely a new plan that requires a lot of changes in the decisions, which is not preferable by manufacturing firms. Therefore, in this research whenever a disruption occurs or a new order arrives, the proposed approach reschedules the remaining manufacturing operations in such a way that minimum changes occur in the original manufacturing plan. Evolutionary optimisation methods have been quite successful and widely addressed by researchers to handle such complex multi-objective optimisation problems because of their ability to find multiple optimal solutions in one single simulation run. Inspired by this, the present research proposes a multiple ant colony optimisation (MACO) algorithm to resolve the computational complexity of a manufacturing rescheduling problem. The performance of the proposed MACO algorithm will be compared with the simple ant colony optimisation (ACO) to judge its robustness and efficacy.

Highly turbulent environment of dynamic job-shop operations affects shop-floor layout as well as manufacturing operations. Due to the dynamic nature of layout changes, essential requirements such as adaptability and responsiveness to the changes need to be considered in addition to the cost issues of material handling and machine relocation when reconfiguring a shop floor’s layout. Here, based on the source of uncertainty, the shop-floor layout problem is split into two sub-problems and dealt with by two modules: re-layout and find-route. Genetic algorithm is used where changes cause the entire shop re-layout, while function blocks are utilised to find the best sequence of robots for the new conditions within the existing layout. This chapter reports the latest development to the authors’ previous work.

Port designers are facing challenges in choosing appropriate terminal layouts to maximise operational efficiencies. This study aims to address this problem by providing a simulation optimisation framework for container terminal layout design. This framework consists of three main modules which are automated layout generator (ALG), the multi-objective optimal computing budget allocation (MOCBA) algorithm and the genetic algorithm (GA). ALG is to automatically generate a simulation model for a set of given design parameters; MOCBA is to intelligently determine the simulation replications to different designs for identifying promising designs; GA is to help generate new design parameters for optimisation. Numerical examples are used to demonstrate the applicability of this framework.

This chapter introduces a novel methodology for the analysis and optimization of production systems. The methodology is based on the innovization procedure, originally introduced for unveiling new and innovative design principles in engineering design problems. Although the innovization method is based on multi-objective optimization and post-optimality analyses of optimised solutions, it stretches the scope beyond an optimization task and attempts to discover new design/operational rules/principles relating to decision variables and objectives, so that a deeper understanding of the problem can be obtained. By integrating the concept of innovization with discrete-event simulation and data mining techniques, a new set of powerful tools can be developed for general systems analysis, particularly suitable for production systems. The uniqueness of the integrated approach proposed in this chapter lies on applying data mining to the data sets generated from simulation-based multi-objective optimization, in order to automatically or semi-automatically discover and interpret the hidden relationships and patterns for optimal production systems design/reconfiguration. After describing the simulation-based innovization using data mining procedure and its difference from conventional simulation analysis methods, results from an industrial case study carried out for the improvement of an assembly line in an automotive manufacturer will be presented.

In recent years simulation-based multi-objective optimisation (SMO) of production systems targeting e.g., throughput, buffers and work-in-process (WIP) has been proven to be a very promising concept. In combination with post-optimality analysis, the concept has the potential of creating a foundation for decision support. This chapter will explore the possibility to expand the concept of introducing optimisation of production system cost aspects such as investments and running cost. A method with a procedure for industrial implementation is presented, including functions for running cost estimation and investment combination optimisation. The potential of applying SMO and post-optimality analysis, taking into account both productivity and financial factors for decision-making support, has been explored and proven to be very beneficial for this kind of industrial application. Evaluating several combined minor improvements with the help of SMO has opened the opportunity to identify a set of solutions (designs) with great financial improvement, which are not feasible to be explored by using current industrial procedures.

This chapter addresses the design of supply chain networks including both network configuration and related operational decisions such as order splitting, transportation allocation and inventory control. The goal is to achieve the best compromise between cost and customer service level. An optimisation methodology that combines a multi-objective genetic algorithm (MOGA) and simulation is proposed to optimise not only the structure of the network but also its operation strategies and related control parameters. A flexible simulation framework is developed to enable the automatic simulation of the supply chain network with all possible configurations and all possible control strategies. To illustrate its effectiveness, the proposed methodology is applied to two real-life case studies from automotive industry and textile industries.