Saeed Zolfaghari

Chaotic time series prediction with residual analysis method using hybrid Elman-NARX neural networks

Residual analysis using hybrid Elman-NARX neural network along with embedding theorem is used to analyze and predict chaotic time series. Using embedding theorem, the embedding parameters are determined and the time series is reconstructed into proper phase space points. The embedded phase space points are fed into an Elman neural network and trained. The residual of predicted time series is analyzed, and it was observed that residuals demonstrate chaotic behaviour. The residuals are considered as a new chaotic time series and reconstructed according to embedding theorem. A new Elman neural network is trained to predict the future value of the residual time series. The residual analysis is repeated several times. Finally, a NARX network is used to capture the relationship among the predicted value of original time series and residuals and original time series. The method is applied to Mackey-Glass and Lorenz equations which produce chaotic time series, and to a real life chaotic time series, Sunspot time series, to evaluate the validity of the proposed technique. Numerical experimental results confirm that the proposed method can predict the chaotic time series more effectively and accurately when compared with the existing prediction methods.

Adaptive temperature control for simulated annealing: a comparative study

In this paper, two variations of simulated annealing method have been proposed and tested on the minimum makespan job shop scheduling problems. In the conventional simulated annealing, the temperature declines constantly, providing the search with a higher transition probability in the beginning of the search and lower probability toward the end of the search. In the first proposed method, an adaptive temperature control scheme is used that changes temperature based on the number of consecutive improving moves. In the second method, a tabu list has been added to the adaptive simulated annealing algorithm in order to avoid revisits. The performance of these two algorithms is evaluated and favorably compared with the conventional simulated annealing.

Erratum to A new genetic algorithm for the machine/part grouping problem involving processing times and lot sizes

This paper reports a new genetic algorithm (GA) for solving a general machine/part grouping (GMPG) problem. In the GMPG problem, processing times, lot sizes and machine capacities are all explicitly considered. To evaluate the solution quality of this type of grouping problems, a generalized grouping efficacy index is used as the performance measure and fitness function of the proposed genetic algorithm. The algorithm has been applied to solving several well-cited problems with randomly assigned processing times to all the operations. To examine the effects of the four major factors, namely parent selection, population size, mutation rate, and crossover points, a large grouping problem with 50 machines and 150 parts has been generated. A multi-factor (34) experimental analysis has been earned out based on 324 GA solutions. The multi-factor ANOVA test results clearly indicate that all the four factors have a significant effect on the grouping output. It is also shown that the interactions between most of the four factors are significant and hence their cross effects on the solution should be also considered in solving GMPG problems.

A COMPARATIVE STUDY OF A NEW HEURISTIC BASED ON ADAPTIVE MEMORY PROGRAMMING AND SIMULATED ANNEALING: THE CASE OF JOB SHOP SCHEDULING

In this study, a general framework is proposed that combines the distinctive features of three well-known approaches: the adaptive memory programming, the simulated annealing, and the tabu search methods. Four variants of a heuristic based on this framework are developed and presented. The performance of the proposed methods is evaluated and compared with a conventional simulated annealing approach using benchmark problems for job shop scheduling. The unique feature of the proposed framework is the use of two short-term memories. The first memory temporarily prevents further changes in the configuration of a provisional solution by maintaining the presence of good elements of such solutions. The purpose of the second memory is to keep track of good solutions found during an iteration, so that the best of these can be used as the starting point in a subsequent iteration. Our computational results for the job shop scheduling problem clearly indicate that the proposed methods significantly outperform the conventional simulated annealing.

Production planning for a ramp-up process with learning in production and growth in demand

This paper presents a production-planning model for a manufacturing process that undergoes a ramp-up period with learning in production and growth in demand. The labour production and demand functions assumed in this paper are validated using available empirical data. A mathematical programming model is developed with numerical examples presented. The results of the paper indicate that the total costs of production can be minimised if the facility produces without interruption during the ramp-up phase and if the production and demand rates are synchronised as much as possible. The latter can be achieved by producing with the lowest possible production rate and by frequently re-structuring the workforce assigned to the production line.

RESIDUAL ANALYSIS AND COMBINATION OF EMBEDDING THEOREM AND ARTIFICIAL INTELLIGENCE IN CHAOTIC TIME SERIES FORECASTING

A combination of embedding theorem and artificial intelligence along with residual analysis is used to analyze and forecast chaotic time series. Based on embedding theorem, the time series is reconstructed into proper phase space points and fed into a neural network whose weights and biases are improved using genetic algorithms. As the residuals of predicted time series demonstrated chaotic behavior, they are reconstructed as a new chaotic time series. A new neural network is trained to forecast future values of residual time series. The residual analysis is repeated several times. Finally, a neural network is trained to capture the relationship among the predicted value of the original time series, residuals, and the original time series. The method is applied to two chaotic time series, Mackey-Glass and Lorenz, for validation, and it is concluded that the proposed method can forecast the chaotic time series more effectively and accurately than existing methods.

Quantitative Models for Centralised Supply Chain Coordination

The effectiveness of coordina tion in supply chains could be measured in two ways: reduction in total supply chain costs and enhanced coordination services provided to the end customer ⎯ and to all players in the supply chain. Inventory is the highest cost in a supply chain accounting for almost 50% of the total logistics costs. Integrating order quantities models among players in a supply chain is a method of achieving coordination. For coordination to be successful, incentive schemes must be adopted. The literature on supply chain coordination have proposed several incentive schemes for coordination; such as quantity discounts, permissible delay in payments, price discounts, volume discount, common replenishment periods. The available quantitative models in supply chain coordination consider up to four levels (i.e., tier-1 supplier, tier-2 supplier, manufacturer, and buyer), with the majority of studies investigating a two-level supply chain with varying assumptions (e.g., multiple buyers, stochastic demand, imperfect quality, etc). Coordination decisions in supply chains are either centralized or decentralized decision-making processes. A centralized decision making process assumes a unique decision-maker (a team) managing the whole supply chain with an objective to minimize (maximize) the total supply chain cost (profit), whereas a decentralized decision-making process involves multiple decision-makers who have conflicting objectives. This chapter will review the literature for quantitative models for centralised supply chain coordination that emphasize inventory management for the period from 1990 to end of 2007. In this chapter, we will classify the models on the basis of incentive schemes, supply chain levels, and assumptions. This chapter will also provide a map indicative of the limitations of the available studies and steer readers to future directions along this line of research.

Supply Uncertainty and Diversification: A Review

We review inventory models that use multiple sourcing (diversification) to deal with upstream (supply) uncertainty. To provide a structured review, we identify three sources of supply uncertainty as follows: supply timing, supply quantity (or quality), and purchase price. Then, we summarize the main results that exist in the literature. Finally, based on our observations, we provide directions for future research.

A virtual collaborative maintenance architecture for manufacturing enterprises

This paper presents a virtual collaborative maintenance architecture aimed at improving the performance of manufacturing systems. The proposed architecture incorporates maintenance elements such as operational reliability, maintenance economics, human factors in maintenance, maintenance program, and maintenance optimization in a virtual collaborative architecture. An analytical model is proposed to measure the relative performance of the proposed virtual collaborative architecture as well as that of the manufacturing enterprise. A numerical example is also presented to demonstrate the application of the proposed approach.

Comprehensive machine cell/part family formation using genetic algorithms

The solution quality of a comprehensive machine/part grouping problem, where the processing times, lot sizes and machine capacities are considered, may not be properly evaluated using a binary performance measure. This paper suggests a generalized grouping efficacy index which has been compared favorably with two binary performance measures. A genetic algorithm using the generalized performance measure as the objective is developed to solve the comprehensive grouping problems. The algorithm has been tested using a number of reference problems with processing times being randomly assigned to all operations. The effects of three major genetic parameters (population size, mutation rate and the number of crossover points) have also been examined. The results indicate that, when the computational time is fixed, larger population size and lower mutation rate tend to improve solution quality while the number of crossover points has no significant impact on the final solution.