Pan Zou

A split-optimization approach for obtaining multiple solutions in single-objective process parameter optimization

Abstract It can be observed from the experimental data of different processes that different process parameter combinations can lead to the same performance indicators, but during the optimization of process parameters, using current techniques, only one of these combinations can be found when a given objective function is specified. The combination of process parameters obtained after optimization may not always be applicable in actual production or may lead to undesired experimental conditions. In this paper, a split-optimization approach is proposed for obtaining multiple solutions in a single-objective process parameter optimization problem. This is accomplished by splitting the original search space into smaller sub-search spaces and using GA in each sub-search space to optimize the process parameters. Two different methods, i.e., cluster centers and hill and valley splitting strategy, were used to split the original search space, and their efficiency was measured against a method in which the original search space is split into equal smaller sub-search spaces. The proposed approach was used to obtain multiple optimal process parameter combinations for electrochemical micro-machining. The result obtained from the case study showed that the cluster centers and hill and valley splitting strategies were more efficient in splitting the original search space than the method in which the original search space is divided into smaller equal sub-search spaces.

A new algorithm based on evolutionary computation for hierarchically coupled constraint optimization: methodology and application to assembly job-shop scheduling

Hierarchically coupled constraint optimization (HCCO) problems are omnipresent, both in theoretical problems and in real-life scenarios; however, there is no clear definition to identify these problems. Numerous techniques have been developed for some typical HCCO problems, such as assembly job-shop scheduling problems (AJSSPs); however, these techniques are not universally applicable to all HCCO problems. In this paper, an abstract definition and common principles amongst different HCCO problems are first established. Next, based on the definitions and principles, a new optimization algorithm based on evolutionary computation is developed for HCCO. The new optimization algorithm has three key new features: a new initial solution generator, a level barrier-based crossover operator, and a level barrier-based mutation operator. In the initial solution generator, a partial solution is created in the first step that satisfies the lowest level hierarchically coupled constraint (HCC) and each consecutive step afterwards adds on to the partial solution to satisfy the next higher level of HCC. In the level barrier-based operators, the operations are only performed between genes satisfying the same level of HCCs to ensure feasibility of the new solutions. The developed optimization algorithm is used to solve a variety of AJSSPs and the results obtained using the proposed algorithm are compared to other methods used to solve AJSSPs.

A Hybrid RF-GA Approach to Bottleneck Machine Diagnosis and Suggestion in Parallel Machine Job-Shop Scheduling

In this paper, a hybrid Random Forest-Genetic Algorithm approach to detect and solve bottleneck machine problems in parallel machine Job-shop scheduling is developed with the aim of minimizing the makespan and the additional cost. The drawbacks of the existing methods for diagnosing bottlenecks is that they either do not consider the severity of the bottleneck or they do not consider the existence of multiple bottlenecks. In the existing models for solving bottlenecks, the cost is not considered as an objective function and only shifting of bottlenecks is utilized to solve the bottleneck machine problem. This approach is not feasible if the maximum capacity of the workshop has been reached. In this paper, a Random Forest classification model is utilized to diagnose bottleneck machine with different severity where the severity of the machines on the shop floor can either be none, low, medium, or high. Due to the lack of historical data, the Random Forest algorithm is trained using bottleneck data generated by simulating several identical parallel machine Job-shop scheduling problems. The trained Random Forest algorithm is then used in conjunction with Genetic Algorithm for finding the optimal actions to be taken for the most severe bottlenecks machines in order to reduce the makespan and the additional cost by optimizing the number of additional parallel machines to be utilized and overtime hours for the most severe bottleneck machines. The two objectives, makespan and additional cost, are combined into a single objective value by the use of weight values. These weight values depend on severity of the most severe bottleneck machine. If the bottleneck severity is “high” then makespan has a higher weight value than cost, if the severity is “medium” then both cost and makespan are weighed equally, and if the severity is “low” then cost has a higher weight value than makespan. In order to show the validity of the proposed approach it is used for diagnosing and solving the bottleneck problems in three different identical parallel machine Job-shop scheduling case studies 1. 3 jobs with 6 machines 2. 5 jobs with 9 machines and 3. 5 jobs with 12 machines. By utilizing the proposed approach the makespan and cost were reduced by 19.0%, 24.5% and 25.4% in case studies 1, 2, and 3 respectively. The results show that the trained Random Forest algorithm was able to correctly diagnose the bottleneck machines and their severity and Genetic algorithm was able to find the optimal number of additional hours and additional machines for the most severe bottleneck machines on the shop floor.© 2016 ASME

An Improved General Regression Neural Network for Prediction Based on Small Samples Data

This paper proposed an improved General Regression Neural Network (GRNN) for prediction based on small samples data by adding the procedure of filtrating the input variables, since the training of original GRNN relies too heavily on data samples and is lack of the relevant process to deal with the errors, while the measurement error and sample error from a plenty of input and output variables in small samples data cannot be easily recognized, but could obviously influence the effect of the training. All the input variables were divided into critical factors and non-critical factors by the impact degree of every input variable, in accordance with the result of the partial correlation analysis on the experimental data. The critical factors were considered as the input of the GRNN model while the non-critical factors were removed to eliminate the error. To verify the performance of the proposed approach, a test on the case of residual stress prediction based on fourteen groups of experimental data was taken as an example application. The improved method could not only avoid the limit to the orthogonality and the quantity of the experiments in traditional method of residual stress prediction, but also get prediction results with smaller error, compared with the original GRNN method. The results have shown that the improved GRNN outperformed the original algorithm with average 4.1% lower prediction error in 92% of the tested cases. And in the rest 8% of tested cases, the prediction error of the improved GRNN was only 0.24% higher than the original one’s.

A Hybrid Neural Network for Prediction of Surface Roughness in Machining

Surface roughness is an important outcome in the machining process and it plays a major role in the manufacturing system. Prediction of surface roughness has been a challenge to researchers because it is impacted by different machining parameters and the inherent uncertainties in the machining process. Prediction of surface roughness will benefit the manufacturing process to be more productive and competitive at the same time to reduce any pre-processing of the machined workpiece in order to meet the technical specifications. In this study, a hybrid GA-LM ANN is proposed for the prediction of surface roughness during roughing process in turning operation. To verify the performance of the proposed approach, the results are compared with the results obtained by training an ANN using GA or LM. The results have shown that the hybrid ANN outperformed the other two algorithms.