Peigen Li

A very fast TS/SA algorithm for the job shop scheduling problem

The job shop scheduling problem (JSP) is one of the most notoriously intractable NP-complete optimization problems. Over the last 10-15 years, tabu search (TS) has emerged as an effective algorithmic approach for the JSP. However, the quality of solutions found by tabu search approach depends on the initial solution. To overcome this problem and provide a robust and efficient methodology for the JSP, the heuristics search approach combining simulated annealing (SA) and TS strategy is developed. The main principle of this approach is that SA is used to find the elite solutions inside big valley (BV) so that TS can re-intensify search from the promising solutions. This hybrid algorithm is tested on the standard benchmark sets and compared with the other approaches. The computational results show that the proposed algorithm could obtain the high-quality solutions within reasonable computing times. For example, 17 new upper bounds among the unsolved problems are found in a short time.

A tabu search algorithm with a new neighborhood structure for the job shop scheduling problem

Tabu search (TS) algorithms are among the most effective approaches for solving the job shop scheduling problem (JSP) which is one of the most difficult NP-complete problems. However, neighborhood structures and move evaluation strategies play the central role in the effectiveness and efficiency of the tabu search for the JSP. In this paper, a new enhanced neighborhood structure is proposed and applied to solving the job shop scheduling problem by TS approach. Using this new neighborhood structure combined with the appropriate move evaluation strategy and parameters, we tested the TS approach on a set of standard benchmark instances and found a large number of better upper bounds among the unsolved instances. The computational results show that for the rectangular problem our approach dominates all others in terms of both solution quality and performance.

Integrating data mining and rough set for customer group-based discovery of product configuration rules

Product configuration design is of critical importance in design for mass customization. This paper will investigate two important issues in configuration design. The first issue is requirement configuration and a dependency analysis approach is proposed and implemented to link customer groups with clusters of product specifications. The second issue concerns the engineering configuration and it is modelled as an association relation between clusters of product specifications and configuration alternatives. A novel methodology and architecture are proposed for accomplishing the two configuration tasks and bridging the gap between them. This methodology is based on integration of popular data mining approaches (such as fuzzy clustering and association rule mining) and variable precision rough set. It focuses on the discovery of configuration rules from the purchased products according to customer groups. The proposed methodology is illustrated with a case study of an electrical bicycle.

A new hybrid GA/SA algorithm for the job shop scheduling problem

Among the modern heuristic methods, simulated annealing (SA) and genetic algorithms (GA) represent powerful combinatorial optimization methods with complementary strengths and weaknesses. Borrowing from the respective advantages of the two paradigms, an effective combination of GA and SA, called Genetic Simulated Algorithm (GASA), is developed to solve the job shop scheduling problem (JSP). This new algorithm incorporates metropolis acceptance criterion into crossover operator, which could maintain the good characteristics of the previous generation and reduce the disruptive effects of genetic operators. Furthermore, we present two novel features for this algorithm to solve JSP. Firstly, a new full active schedule (FAS) based on the operation-based representation is presented to construct schedule, which can further reduce the search space. Secondly, we propose a new crossover operator, named Precedence Operation Crossover (POX), for the operation-based representation. The approach is tested on a set of standard instances and compared with other approaches. The Simulation results validate the effectiveness of the proposed algorithm.

Agile manufacturing system control based on cell re-configuration

The control of an agile manufacturing system (AMS) is expected to be flexible, open, scalable and re-configurable so as to tackle the more complex and uncertain information flows. To meet these requirements, we propose agent-based control architecture for AMS, under which the functions of task planning, scheduling and dynamic control are integrated seamlessly. First of all, this paper introduces the concept of RMC (re-configurable manufacturing cell), based on which, we construct the control architecture for AMS in compliance with multi-agent system (MAS). The whole control process under the architecture comprises two hierarchies, i.e. the upper one for order planning and RMC forming and the lower one for task scheduling within each RMC. For the upper hierarchy, we establish a linear integer programming (LIP)-based mathematical model and a MAS-based dynamic process model, and present a two-step approach to order planning and RMC forming. For the lower hierarchy, we develop the scheduling model, a method based on the bidding mechanism from contract net, and describe the rescheduling mechanism in the control system. To illustrate the methodology proposed in the paper, a simulation study is thoroughly discussed. Our studies demonstrate that the RMC-based control architecture provides an AMS with an optimal, dynamic and flexible mechanism of responding to an unpredictable manufacturing environment, which is crucial to achieve agility for the whole manufacturing system.

A CORBA- and MAS-based architecture for agile collaborative manufacturing systems

An architecture for building up agile collaborative manufacturing systems (ACMS), which is based on common object request broker architecture (CORBA) and multi-agent system (MAS) paradigm, is proposed in the current paper. First, the diverse manufacturing resources, which may be heterogeneous and physically distributed, are MAS-based agentified and CORBA-based encapsulated as the resource agents. Second, the source agents are registered at the system manager agent (SMA) as member agents and subsequently integrated under the proposed architecture to form an ACMS. Third, the member agents communicate and interact with each other to conduct collaborative manufacturing by means of agent interoperation and human-machine interaction. In order to validate the methodology, an experimental system is developed, where several heterogeneous manufacturing resources are well-integrated via their agents in a plug-and-play manner to establish an ACMS prototype. The experiments demonstrate that the proposed architecture, characterized by distribution and openness, provides a feasible solution for developing ACMS, where integratability, reconfigurability, flexibility and agility are achieved.

Threefold versus fivefold cross-validation and individual versus average data in predictive regression modelling of machining experimental data

Model selection and validation are critical in predicting the performance of manufacturing processes. Proper selection of variables helps minimize the model mismatch error, proper selection of models helps reduce the model estimation error, and proper validation of models helps minimize the model prediction error. In the current paper, the literature is reviewed and a rigorous procedure is proposed for selection and cross-validation (CV) of predictive regression models. Experimental data from a turning surface roughness study are used to demonstrate how to select and validate predictive regression models. In particular, different data splitting methods are compared, such as fivefold CV versus threefold CV as well as the individual data versus the average data. This paper has revealed no statistical difference between the use of fivefold CV and threefold CV, and the use of the individual and the average data in subset selection and CV of predictive regression models. Consequently, threefold instead of fivefold or tenfold CV and either individual data or average data may be used to reduce the computational cost in predictive regression modelling of experimental data based on this and other similar empirical studies.

An approach combined Response Surface Method and Particle Swarm Optimization to ship multidisciplinary design and optimization

Ship design is a complex endeavor requiring the successful coordination of many different disciplines. According to various disciplines requirements, how to get a balanced performance is imperative in ship design. Thus, a all-in-one Multidisciplinary Design Optimization (MDO) approach is proposed to get the optimum performance of the ship considering three disciplines, structure; cargo loads and power of propulsion. In this research a Latin Hypercube Sampling (LHS) is employed to explore the design space and to sample data for covering the design space. For the purpose of reducing the calculation and saving the develop time, a quadratic Response Surface Method (RSM) is adopted as an approximation model for solving the system design problems. Particle Swarm Optimization (PSO) is introduced to search the appropriate design result in MDO in ship design. Finally, the validity of the proposed approach is proven by a case study of a bulk carrier.

A mathematical programming method for flow path design in high-mix and low-volume flow manufacturing

For the operation and control of flow manufacturing in a high-mix and low-volume production environment, a method of flow path management is introduced. It involves the dynamic formation and control of adaptable flow paths corresponding to different product families. A mathematical programming approach is presented for the design of flow paths, which takes into consideration the sharing of machines among various product families. The objective is to maximize the throughput in a predefined planning period and to minimize the number of shared machines. An illustrative example is given for demonstrating the use of the proposed method.

Group Decision-based Collaborative Design

There is a growing recognition that decisions are the fundamental construct in product design. The collaborative design process gets more complex when a group of geographically dispersed designers work as a team. In this paper, we take a group decision-based approach to model design process and introduce an agent-based architecture to support collaborative design. The proposed architecture achieves the design decisions integration between design members in a dynamic design environment. In particular, the proposed agents can provide the interaction spaces for coordinating the individual design decisions. Finally, the mechanism of coordination between agents is discussed