IEEE Transactions on Evolutionary Computation, 2019
Evolutionary algorithms have shown their promise in coping with many-objective optimization probl... more Evolutionary algorithms have shown their promise in coping with many-objective optimization problems. However, the strategies of balancing convergence and diversity and the effectiveness of handling problems with irregular Pareto fronts (PFs) are still far from perfect. To address these issues, this paper proposes an adaptive sorting-based evolutionary algorithm based on the idea of decomposition. First, we propose an adaptive sorting-based environmental selection strategy. Solutions in each subpopulation (partitioned by reference vectors) are sorted based on their convergence. Those with better convergence are further sorted based on their diversity, then being selected according to their sorting levels. Second, we provide an adaptive promising subpopulation sorting-based environmental selection strategy for problems which may have irregular PFs. This strategy provides additional sorting-based selection effort on promising subpopulations after the general environmental selection process. Third, we extend the algorithm to handle constraints. Finally, we conduct an extensive experimental study on the proposed algorithm by comparing with start-of-the-state algorithms. Results demonstrate the superiority of the proposed algorithm.
Determining optimal bidding strategies in a competitive electricity market to maximize the profit... more Determining optimal bidding strategies in a competitive electricity market to maximize the profit of each bidder is a challenging economic game problem. In this paper, it is formulated as a bi-level optimization problem in which, in the lower level, the community's social welfare is maximized by solving a power flow problem while, in the upper level, the profits of individual bidders are maximized. In this bidders' game, instead of using a set of discrete strategies as is usual, we consider continuous functions as strategies. To solve the upper-level problem, two co-evolutionary approaches are proposed and, for the lower level, an interior point algorithm is applied. Three IEEE benchmark problems in four different scenarios are solved and their results compared with those obtained from two conventional approaches and the literature which indicate that the proposed approaches have some merit regarding quality and efficiency.
Capability-based planning (CBP) is a strategy focused planning framework that facilitates organiz... more Capability-based planning (CBP) is a strategy focused planning framework that facilitates organizations to systematically develop capacity to achieve their business objectives in highly uncertain, dynamic and competitive environments. Capability programming is an integral part of CBP which requires selecting a portfolio of capability projects for execution, referred as a capability program, such that the overall strategic risk facing the planning organization across a number of projected future operating scenarios is minimized while maintaining the most economical choice. It is a challenging optimization problem that requires handling a number of dynamic constraints and objectives that vary throughout the entire planning horizon. An optimizing simulation approach is presented in this paper that combines an evolutionary multi-objective optimization algorithm with a reinforcement learning technique to generate capability programs which optimize strategic risks and program costs across multiple planning scenarios as well as over a rolling planning horizon. The role of the optimization algorithm in this approach is to search for the non-dominated capability programs at each decision point by minimizing the strategic risks associated with individual capability projects across a number of planning scenarios as well as the total cost of the program. The reinforcement learning algorithm, on the other hand, searches horizontally within the set of non-dominated programs to minimize capability risks and costs over the entire planning horizon. The methodology is evaluated on a test problem generated based on the data distributions in an Australian Defence Capability Plan and the performance is compared with two myopic heuristic methods.
Investigating Multi-Operator Differential Evolution for Feature Selection
Lecture Notes in Computer Science, 2016
Performance issues when dealing with a large number of features are well-known for classification... more Performance issues when dealing with a large number of features are well-known for classification algorithms. Feature selection aims at mitigating these issues by reducing the number of features in the data. Hence, in this paper, a feature selection approach based on a multi-operator differential evolution algorithm is proposed. The algorithm partitions the initial population into a number of sub-populations evolving using a pool of distinct mutation strategies. Periodically, the sub-populations exchange information to enhance their diversity. This multi-operator approach reduces the sensitivity of the standard differential evolution to the selection of an appropriate mutation strategy. Two classifiers, namely decision trees and k-nearest neighborhood, are used to evaluate the generated subsets of features. Experimental analysis has been conducted on several real data sets using a 10-fold cross validation. The analysis shows that the proposed algorithm successfully determines efficient feature subsets, which can improve the classification accuracy of the classifiers under consideration. The usefulness of the proposed method on large scale data set has been demonstrated using the KDD Cup 1999 intrusion data set, where the proposed method can effectively remove irrelevant features from the data.
2014 IEEE International Conference on Industrial Engineering and Engineering Management, 2014
For successful operation of any power system, an effective scheduling of power generation is cruc... more For successful operation of any power system, an effective scheduling of power generation is crucial. In this paper, we consider a power system with two types of generators, thermal and hydro. The characteristics of these generators vary with respect to the cost, emission to the environment, input source, capacity limit, and technological constraints. The mathematical model considering two objectives, such as minimization of the operating cost and minimization of total emissions, for a hydrothermal system is discussed. A solution approach has been proposed, based on evolutionary computation concept, for solving a benchmark problem for both single and bi-objective version of the problem. In the approach, an initial population of solutions is generated based on a heuristic and the population is then evolved using two well-known evolutionary search algorithms. The solutions of our approaches are compared with another approach from the literature. The analysis of the results reveals that the heuristic enhanced the performance of the evolutionary algorithms considered in this paper.
Students recruited group 1 Students recruited group 2 Students recruited group 3 Students recruit... more Students recruited group 1 Students recruited group 2 Students recruited group 3 Students recruited group 4 Students recruited group 5 Eligible students in group 1 Eligible students in group 2 Eligible students in group 3 Eligible students in group 4 Eligible students in group 5
2015 IEEE Congress on Evolutionary Computation (CEC), 2015
The resource constrained project scheduling problem (RCPSP) has a wide variety of practical appli... more The resource constrained project scheduling problem (RCPSP) has a wide variety of practical applications in construction, manufacturing, project planning, and other areas. Since the 1960s many optimization algorithms have been proposed to solve this NP-hard problem, and their performances are evaluated in well-known test problems with different complexities. Although it is desirable to find an algorithm which can provide promising solutions with reasonable computational efforts for any problem under consideration, no single algorithm can meet that condition. To deal with this challenge, we present a genetic algorithm based memetic algorithm (MA) for solving RCPSP. The algorithm is initiated by a critical path-based heuristic and a variant of the Nawaz, Enscore, and Ham (NEH) heuristic. The algorithm involves a similar block order crossover and a variable insertion based local search. An automatic restart scheme is also presented which assists the algorithm to escape from local optima. In addition, a design-of-experiment (DOE) method is used to determine the set of suitable parameters for the proposed MA. Numerical results, statistical analysis and comparisons with state-of-the-art algorithms demonstrate the effectiveness of the proposed approach. 2 1 j (4) a {0, 1} jt (5) The objective function is the minimization of makespan, C max (Eq. (1)). Eq. (2) ensures that an activity can be executed only once. Eq. (3) ensures that each activity j cannot be started unless all its predecessors have been completed. Eq. (4) ensures that an activity can be started when its required renewable resources (such as workforce, machines, tools or equipment) are available. Over the years, exact techniques such as branch and bound [4-6], branch and cut [7], and the event based approach [8] have been proposed for the optimal solution of RCPSPs. Due to computational
In this paper, a multi-echelon, multi-period, decentralized supply chain (SC) with a single manuf... more In this paper, a multi-echelon, multi-period, decentralized supply chain (SC) with a single manufacturer, single distributor and single retailer is considered. For this setting, a twophase planning approach combining centralized and decentralized decision-making processes is proposed, in which the first-phase planning is a coordinated centralized controlled, and the second-phase planning is viewed as independent decentralized decision-making for individual entities. This research focuses on the independence and equally powerful behavior of the individual entities with the aim of achieving the maximum profit for each stage. A mathematical model for total SC coordination as a first-phase planning problem and separate ones for each of the independent members with their individual objectives and constraints as second-phase planning problems are developed. We introduce a new solution approach using a goal programming technique in which a target or goal value is set for each independent decision problem to ensure that it obtains a near value for its individual optimum profit, with a numerical analysis presented to explain the results. Moreover, the proposed two-phase model is compared with a single-phase approach in which all stages are considered dependent on each other as parts of a centralized SC. The results prove that the combined two-phase planning method for a decentralized SC network is more realistic and effective than a traditional single-phase one.
This paper presents the study of a real-time procurement and production mechanism for a three-sta... more This paper presents the study of a real-time procurement and production mechanism for a three-stage supply chain system with multiple suppliers, subject to unexpected disruptions. In the first part of the paper, a mathematical model was developed for the optimization of replenishment and production decisions for each node after the occurrence of a transportation disruption. In addition, an experiment was conducted to study the effects of disruptions to the system using predefined scenarios, where the supplier’s prioritization of disruption mitigation strategies was explored. Various disruption scenarios were predefined by combining different disruption types and locations as well as different combinations of suppliers. It will be shown that the solution to the transportation disruption was more sensitive to the lot size when the lost sales cost was large. However, when the lost sales cost was low, the sensitivity to the lot size decreased, and the setup cost and inventory holding co...
The performance of a Convolutional Neural Network (CNN) highly depends on its architecture and co... more The performance of a Convolutional Neural Network (CNN) highly depends on its architecture and corresponding parameters. Manually designing a CNN is a time-consuming process in regards to the various layers that it can have, and the variety of parameters that must be set up. Increasing the complexity of the network structure by employing various types of connections makes designing a network even more challenging. Evolutionary computation as an optimisation technique can be applied to arrange the CNN layers and/or initiate its parameters automatically or semi-automatically. Dense network and Residual network are two popular network structures that were introduced to facilitate the training of deep networks. In this paper, leveraging the potentials of Dense and Residual blocks, and using the capability of evolutionary computation, we propose an automatic evolutionary model to detect an optimum and accurate network structure and its parameters for medical image segmentation. The proposed evolutionary DenseRes model is employed for segmentation of six publicly available MRI and CT medical datasets. The proposed model obtained high accuracy while employing networks with minimal parameters for the segmentation of medical images and outperformed manual and automatic designed networks, including U
2014 IEEE International Conference on Industrial Engineering and Engineering Management, 2014
In this paper, a supply disruption management model is introduced in a three-tier supply chain wi... more In this paper, a supply disruption management model is introduced in a three-tier supply chain with multiple suppliers and retailers, where the system may face sudden disruption in its raw material supply. At first, we formulated a mathematical model for ideal conditions and then reformulated it to revise the supply, production and delivery plan after the occurrence of a disruption, for a future period, to recover from the disruption. Here, the objective is to minimize the total cost during the recovery time window while being subject to supply, capacity, demand, and delivery constraints. We have also proposed an efficient heuristic to solve the model and the results have been compared, with another established solution approach, for a good number of randomly generated test problems. The comparison showed the consistent performance of our developed heuristic. This paper also presents some numerical examples to explain the usefulness of the proposed approach.
Supply chains face risks from various unexpected events that make disruptions almost inevitable. ... more Supply chains face risks from various unexpected events that make disruptions almost inevitable. This paper presents a disruption recovery model for a single stage production and inventory system, where finished product supply is randomly disrupted for periods of random duration. A production facility that manufactures a single product following the Economic Production Quantity policy is considered. The model is solved using a search algorithm combined with a penalty function method to find the best recovery plan. It is shown that the optimal recovery schedule is dependent on the extent of the disruption, as well as the back order cost and lost sales cost parameters. The proposed model is seen to be a very useful tool for manufacturers to make quick decisions on the optimal recovery plan after the occurrence of a disruption.
Journal of Industrial and Management Optimization, 2015
This paper presents a literature review on risk and disruption management in productioninventory ... more This paper presents a literature review on risk and disruption management in productioninventory and supply chain systems. The review is conducted on the basis of comparing various works published in this research domain, specifically the papers, which considered real-life risk factors, such as imperfect production processes, risk and disruption in production, supply, demand, and transportation, while developing models for productioninventory and supply chain systems. Emphasis is given on the assumptions and the types of problems considered in the published research. We also focus on reviewing the mathematical models and the solution approaches used in solving the models using both hypothetical and real-world problem scenarios. Finally, the literature review is summarized and future research directions are discussed.
Managing risk in production scheduling under uncertain disruption
Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 2015
The job scheduling problem (JSP) is considered as one of the most complex combinatorial optimizat... more The job scheduling problem (JSP) is considered as one of the most complex combinatorial optimization problems. JSP is not an independent task, but is rather a part of a company business case. In this paper, we have studied JSPs under sudden machine breakdown scenarios that introduce a risk of not completing the jobs on time. We have first solved JSPs using an improved memetic algorithm and extended the algorithm to deal with the disruption situations, and then developed a simulation model to analyze the risk of using a job order and delivery scenario. This paper deals with job scheduling under ideal conditions and rescheduling under machine breakdown, and provides a risk analysis for a production business case. The extended algorithm provides better understanding and results than existing algorithms, the rescheduling shows a good way of recovering from disruptions, and the risk analysis shows an effective way of maximizing return under such situations.
A real-time order acceptance and scheduling approach for permutation flow shop problems
European Journal of Operational Research, 2015
ABSTRACT The Permutation Flow Shop Scheduling Problem (PFSP) is a complex combinatorial optimizat... more ABSTRACT The Permutation Flow Shop Scheduling Problem (PFSP) is a complex combinatorial optimization problem. PFSP has been widely studied as a static problem using heuristics and metaheuristics. In reality, PFSPs are not usually static, but are rather dynamic, as customer orders are placed at random time intervals. In the dynamic problem, two tasks must be considered: (i) should a new order be accepted? and (ii) if accepted, how can this schedule be ordered, when some orders may be already under process and or be in the queue for processing? For the first task, we propose a simple heuristic based decision process, and for the second task, we developed a Genetic Algorithm (GA) based approach that is applied repeatedly for re-optimization as each new order arrives. The usefulness of the proposed approach has been demonstrated by solving a set of test problems. In addition the proposed approach, along with a simulation model, has been tested for maximizing the revenue of a flow shop production business under different order arrival scenarios. Finally, a case study is presented to show the applicability of the proposed approach in practice.
Neurodynamic differential evolution algorithm and solving CEC2015 competition problems
2015 IEEE Congress on Evolutionary Computation (CEC), 2015
ABSTRACT Recently, the success history based parameter adaptation for differential evolution algo... more ABSTRACT Recently, the success history based parameter adaptation for differential evolution algorithm with linear population size reduction has been claimed to be a great algorithm for solving optimization problems. Neuro-dynamic is another recent approach that has shown remarkable convergence for certain problems, even for high dimensional cases. In this paper, we proposed a new algorithm by embedding the concept of neuro-dynamic into a modified success history based parameter adaptation for differential evolution with linear population size reduction. We have also proposed an adaptive mechanism for the appropriate use of the success history based parameter adaptation for differential evolution with linear population size reduction and neuro-dynamic during the search process. The new algorithm has been tested on the CEC’2015 single objective real-parameter competition problems. The experimental results show that the proposed algorithm is capable of producing good solutions that are clearly better than those obtained from the success history based parameter adaptation for differential evolution with linear population size reduction and a few of the other state-of-the-art algorithms considered in this paper
Ieee Congress on Evolutionary Computation, Jul 18, 2010
The job scheduling problem (JSP) is considered as one of the complex combinatorial optimization p... more The job scheduling problem (JSP) is considered as one of the complex combinatorial optimization problems. In this paper, we have developed a hybrid Genetic Algorithm (HGA), which improves the performance of GAs when solving JSPs. We have also modified the developed algorithm to study JSPs under the machine unavailability condition. We have considered two types of machine unavailability. Firstly, where the unavailability information is available in advance (predictive) and, secondly, where the information is known after a real breakdown (reactive). We have shown that the revised schedule is mostly able to recover if the disruptions occur during the early stages of a schedule.
The number of just-in-time jobs maximization in a permutation flow shop scheduling problem is con... more The number of just-in-time jobs maximization in a permutation flow shop scheduling problem is considered. A mixed integer linear programming model to represent the problem as well as solution approaches based on enumeration and constructive heuristics were proposed and computationally implemented. Instances with up to 10 jobs and five machines are solved by the mathematical model in an acceptable running time (3.3 min on average) while the enumeration method consumes, on average, 1.5 s. The 10 constructive heuristics proposed show they are practical especially for large-scale instances (up to 100 jobs and 20 machines), with very good-quality results and efficient running times. The best two heuristics obtain near-optimal solutions, with only 0.6% and 0.8% average relative deviations. They prove to be better than adaptations of the NEH heuristic (well-known for providing very good solutions for makespan minimization in flow shop) for the considered problem.
Uploads
Papers by Ruhul Sarker