Showing papers on "Job shop scheduling published in 2018"

Optimus: an efficient dynamic resource scheduler for deep learning clusters

Abstract: Deep learning workloads are common in today's production clusters due to the proliferation of deep learning driven AI services (e.g., speech recognition, machine translation). A deep learning training job is resource-intensive and time-consuming. Efficient resource scheduling is the key to the maximal performance of a deep learning cluster. Existing cluster schedulers are largely not tailored to deep learning jobs, and typically specifying a fixed amount of resources for each job, prohibiting high resource efficiency and job performance. This paper proposes Optimus, a customized job scheduler for deep learning clusters, which minimizes job training time based on online resource-performance models. Optimus uses online fitting to predict model convergence during training, and sets up performance models to accurately estimate training speed as a function of allocated resources in each job. Based on the models, a simple yet effective method is designed and used for dynamically allocating resources and placing deep learning tasks to minimize job completion time. We implement Optimus on top of Kubernetes, a cluster manager for container orchestration, and experiment on a deep learning cluster with 7 CPU servers and 6 GPU servers, running 9 training jobs using the MXNet framework. Results show that Optimus outperforms representative cluster schedulers by about 139% and 63% in terms of job completion time and makespan, respectively.

Learning Scheduling Algorithms for Data Processing Clusters

Abstract: Efficiently scheduling data processing jobs on distributed compute clusters requires complex algorithms. Current systems, however, use simple generalized heuristics and ignore workload characteristics, since developing and tuning a scheduling policy for each workload is infeasible. In this paper, we show that modern machine learning techniques can generate highly-efficient policies automatically. Decima uses reinforcement learning (RL) and neural networks to learn workload-specific scheduling algorithms without any human instruction beyond a high-level objective such as minimizing average job completion time. Off-the-shelf RL techniques, however, cannot handle the complexity and scale of the scheduling problem. To build Decima, we had to develop new representations for jobs' dependency graphs, design scalable RL models, and invent RL training methods for dealing with continuous stochastic job arrivals. Our prototype integration with Spark on a 25-node cluster shows that Decima improves the average job completion time over hand-tuned scheduling heuristics by at least 21%, achieving up to 2x improvement during periods of high cluster load.

An effective and distributed particle swarm optimization algorithm for flexible job-shop scheduling problem

Abstract: Flexible job-shop scheduling problem (FJSP) is very important in many research fields such as production management and combinatorial optimization. The FJSP problems cover two difficulties namely machine assignment problem and operation sequencing problem. In this paper, we apply particle swarm optimization (PSO) algorithm to solve this FJSP problem aiming to minimize the maximum completion time criterion. Various benchmark data taken from literature, varying from Partial FJSP and Total FJSP, are tested. Experimental results proved that the developed PSO is enough effective and efficient to solve the FJSP. Our other objective in this paper, is to study the distribution of the PSO-solving method for future implementation on embedded systems that can make decisions in real time according to the state of resources and any unplanned or unforeseen events. For this aim, two multi-agent based approaches are proposed and compared using different benchmark instances.

Fog Computing for Energy-Aware Load Balancing and Scheduling in Smart Factory

Abstract: Due to the development of modern information technology, the emergence of the fog computing enhances equipment computational power and provides new solutions for traditional industrial applications. Generally, it is impossible to establish a quantitative energy-aware model with a smart meter for load balancing and scheduling optimization in smart factory. With the focus on complex energy consumption problems of manufacturing clusters, this paper proposes an energy-aware load balancing and scheduling (ELBS) method based on fog computing. First, an energy consumption model related to the workload is established on the fog node, and an optimization function aiming at the load balancing of manufacturing cluster is formulated. Then, the improved particle swarm optimization algorithm is used to obtain an optimal solution, and the priority for achieving tasks is built toward the manufacturing cluster. Finally, a multiagent system is introduced to achieve the distributed scheduling of manufacturing cluster. The proposed ELBS method is verified by experiments with candy packing line, and experimental results showed that proposed method provides optimal scheduling and load balancing for the mixing work robots.

Optimal Link Scheduling for Age Minimization in Wireless Systems

Abstract: Information age is a recently introduced metric to represent the freshness of information in communication systems. We investigate age minimization in a wireless network and propose a novel approach of optimizing the scheduling strategy to deliver all messages as fresh as possible. Specifically, we consider a set of links that share a common channel. The transmitter at each link contains a given number of packets with time stamps from an information source that generated them. We address the link transmission scheduling problem with the objective of minimizing the overall age. This minimum age scheduling problem (MASP) is different from minimizing the time or the delay for delivering the packets in question. We model the MASP mathematically and prove it is NP-hard in general. We also identify tractable cases as well as optimality conditions. An integer linear programming formulation is provided for performance benchmarking. Moreover, a steepest age descent algorithm with better scalability is developed. Numerical study shows that, by employing the optimal schedule, the overall age is significantly reduced in comparison to other scheduling strategies.

Residential energy scheduling for variable weather solar energy based on adaptive dynamic programming

Abstract: The residential energy scheduling of solar energy is an important research area of smart grid. On the demand side, factors such as household loads, storage batteries, the outside public utility grid and renewable energy resources, are combined together as a nonlinear, time-varying, indefinite and complex system, which is difficult to manage or optimize. Many nations have already applied the residential real-time pricing to balance the burden on their grid. In order to enhance electricity efficiency of the residential micro grid, this paper presents an action dependent heuristic dynamic programming U+0028 ADHDP U+0029 method to solve the residential energy scheduling problem. The highlights of this paper are listed below. First, the weather-type classification is adopted to establish three types of programming models based on the features of the solar energy. In addition, the priorities of different energy resources are set to reduce the loss of electrical energy transmissions. Second, three ADHDP-based neural networks, which can update themselves during applications, are designed to manage the flows of electricity. Third, simulation results show that the proposed scheduling method has effectively reduced the total electricity cost and improved load balancing process. The comparison with the particle swarm optimization algorithm further proves that the present method has a promising effect on energy management to save cost.

Workflow Scheduling Using Hybrid GA-PSO Algorithm in Cloud Computing

Abstract: Cloud computing environment provides several on-demand services and resource sharing for clients. Business processes are managed using the workflow technology over the cloud, which represents one of the challenges in using the resources in an efficient manner due to the dependencies between the tasks. In this paper, a Hybrid GA-PSO algorithm is proposed to allocate tasks to the resources efficiently. The Hybrid GA-PSO algorithm aims to reduce the makespan and the cost and balance the load of the dependent tasks over the heterogonous resources in cloud computing environments. The experiment results show that the GA-PSO algorithm decreases the total execution time of the workflow tasks, in comparison with GA, PSO, HSGA, WSGA, and MTCT algorithms. Furthermore, it reduces the execution cost. In addition, it improves the load balancing of the workflow application over the available resources. Finally, the obtained results also proved that the proposed algorithm converges to optimal solutions faster and with higher quality compared to other algorithms.

MINLP Probabilistic Scheduling Model for Demand Response Programs Integrated Energy Hubs

Abstract: In this paper, an optimal probabilistic scheduling model of energy hubs operations is presented. The scheduling of energy hub determines the energy carriers to be purchased as input and converted or stored, in order to meet the energy requests, while minimizing the total hub's cost. However, as many other engineering endeavors, future operating criteria of energy hubs could not be forecasted with certainty. Load and price uncertainties are the most unclear parameters that hub operators deal with. In this regard, this paper proposes a 2 $m$ + 1 point estimation probabilistic scheduling scheme for energy hubs with multiple energy inputs and outputs. One of the applicable tools of energy hubs to have an efficient participation in the liberalized power market with volatile prices is demand response programs (DRPs). While there is plenty of experience in investigating the effect of DRP, it is electricity DRP that receives increasing attention by research and industry. Therefore, the proposed DRP investigates the effect of both responsive thermal and electric loads in reducing the total cost and participation of different facilities in supplying multiple loads. The proposed model envisages the most technical constraints of converters and storages. Several test systems have been investigated in order to confirm the effectiveness of the proposed model. The results verify the capability of the proposed model in covering the energy hub time-varying output demands as well as the economic advantages of implementing the suggested strategy. In addition, the results are compared with 2 $m$ point estimate method and Monte Carlo simulation.

Mixed-Integer Linear Programming for Optimal Scheduling of Autonomous Vehicle Intersection Crossing

Abstract: We propose an urban traffic management scheme for an all connected vehicle environment. If all the vehicles are autonomous, for example, in smart city projects or future's dense city centers, then such an environment does not need a physical traffic signal. Instead, an intersection control server processes data streams from approaching vehicles, periodically solves an optimization problem, and assigns to each vehicle an optimal arrival time that ensures safety while significantly reducing number of stops and intersection delays. The scheduling problem is formulated as a mixed-integer linear program (MILP), and is solved by IBM CPLEX optimization package. The optimization outputs (scheduled access/arrival times) are sent to all approaching vehicles. The autonomous vehicles adjust their speed accordingly by a proposed trajectory planning algorithm with the aim of accessing the intersection at their scheduled times. A customized traffic microsimulation environment is developed to determine the potentials of the proposed solution in comparison to two baseline scenarios. In addition, the proposed MILP-based intersection control scheme is modified and simulated for a mixed traffic consisting of autonomous and human-controlled vehicles, all connected through a wireless communication to the intersection controller of a signalized intersection.

Incremental Flow Scheduling and Routing in Time-Sensitive Software-Defined Networks

Abstract: Several networking architectures have been developed atop IEEE 802.3 networks to provide real-time communication guarantees for time-sensitive applications in industrial automation systems. The basic principle underlying these technologies is the precise transmission scheduling of time-triggered traffic through the network for providing deterministic and bounded latency and jitter. These transmission schedules are typically synthesized offline (computational time in the order of hours) and remain fixed thereafter, making it difficult to dynamically add or remove network applications. This paper presents algorithms for incrementally adding time-triggered flows in a time-sensitive software-defined network (TSSDN). The TSSDN is a network architecture based on software-defined networking, which provides real-time guarantees for time-triggered flows by scheduling their transmissions on the hosts (network edge) only. These algorithms exploit the global view of the control plane on the data plane to schedule and route time-triggered flows needed for the dynamic applications in the Industrial Internet of Things (Industry 4.0). The evaluations show that these algorithms can compute incremental schedules for time-triggered flows in subseconds with an average relative optimality of 68%.

A memetic differential evolution algorithm for energy-efficient parallel machine scheduling

Abstract: This paper considers an energy-efficient bi-objective unrelated parallel machine scheduling problem to minimize both makespan and total energy consumption. The parallel machines are speed-scaling. To solve the problem, we propose a memetic differential evolution (MDE) algorithm. Since the problem involves assigning jobs to machines and selecting an appropriate processing speed level for each job, we characterize each individual by two vectors: a job-machine assignment vector and a speed vector. To accelerate the convergence of the algorithm, only the speed vector of each individual evolves and a list scheduling heuristic is applied to derive its job-machine assignment vector based on its speed vector. To further enhance the algorithm, we propose efficient speed adjusting and job-machine swap heuristics and integrate them into the algorithm as a local search approach by an adaptive meta-Lamarckian learning strategy. Computational results reveal that the incorporation of list scheduling heuristic and local search greatly strengthens the algorithm. Computational experiments also show that the proposed MDE algorithm outperforms SPEA-II and NSGA-II significantly.

A novel hybrid multi-objective artificial bee colony algorithm for blocking lot-streaming flow shop scheduling problems

Abstract: A blocking lot-streaming flow shop (BLSFS) scheduling problem is to schedule a number of jobs on more than one machine, where each job is split into a number of sublots while no intermediate buffers exist between adjacent machines. The BLSFS scheduling problem roots from traditional job shop scheduling problems but with additional constraints. It is more difficult to be solved than traditional job shop scheduling problems, yet very popular in real-world applications, and research on the problem has been in its infancy to date. This paper presents a hybrid multi-objective discrete artificial bee colony (HDABC) algorithm for the BLSFS scheduling problem with two conflicting criteria: the makespan and the earliness time. The main contributions of this paper include: (1) developing an initialization approach using a prior knowledge which can produce a number of promising solutions, (2) proposing two crossover operators by taking advantage of valuable information extracted from all the non-dominated solutions in the current population, and (3) presenting an efficient Pareto local search operator based on the Pareto dominance relation. The proposed algorithm is empirically compared with four state-of-the-art multi-objective evolutionary algorithms on 18 test subsets of the BLSFS scheduling problem. The experimental results show that the proposed algorithm significantly outperforms the compared ones in terms of several widely-used performance metrics.

Cooperative Task Scheduling for Computation Offloading in Vehicular Cloud

Abstract: Technological evolutions in the automobile industry, especially the development of connected and autonomous vehicles, have granted vehicles more computing, storage, and sensing resources. The necessity of efficient utilization of these resources leads to the vision of vehicular cloud computing (VCC), which can offload the computing tasks from the edge or remote cloud to enhance the overall efficiency. In this paper, we study the problem of computation offloading through the vehicular cloud (VC), where computing missions from edge cloud can be offloaded and executed cooperatively by vehicles in VC. Specifically, computing missions are further divided into computing tasks with interdependency and executed in different vehicles in the VC to minimize the overall response time. To characterize the instability of computing resources resulting from the high vehicular mobility, a mobility model focusing on vehicular dwell time is utilized. Considering the heterogeneity of vehicular computing capabilities and the interdependency of computing tasks, we formulate an optimization problem for task scheduling, which is NP-hard. For low complexity, a modified genetic algorithm based scheduling scheme is designed where integer coding is used rather than binary coding, and relatives are defined and employed to avoid infeasible solutions. In addition, a task load based stability analysis of the VCC system is presented for the cases where some vehicles within the VC are offline. Numerical results demonstrate that the proposed scheme can significantly improve the utilization of computing resources while guaranteeing low latency and system stability.

Sustainable Service Allocation Using a Metaheuristic Technique in a Fog Server for Industrial Applications

Abstract: Reducing energy consumption in the fog computing environment is both a research and an operational challenge for the current research community and industry. There are several industries such as finance industry or healthcare industry that require a rich resource platform to process big data along with edge computing in fog architecture. As a result, sustainable computing in a fog server plays a key role in fog computing hierarchy. The energy consumption in fog servers depends on the allocation techniques of services (user requests) to a set of virtual machines (VMs). This service request allocation in a fog computing environment is a nondeterministic polynomial-time hard problem. In this paper, the scheduling of service requests to VMs is presented as a bi-objective minimization problem, where a tradeoff is maintained between the energy consumption and makespan. Specifically, this paper proposes a metaheuristic-based service allocation framework using three metaheuristic techniques, such as particle swarm optimization (PSO), binary PSO, and bat algorithm. These proposed techniques allow us to deal with the heterogeneity of resources in the fog computing environment. This paper has validated the performance of these metaheuristic-based service allocation algorithms by conducting a set of rigorous evaluations.

Effect of Natural Gas Flow Dynamics in Robust Generation Scheduling Under Wind Uncertainty

Abstract: A new challenge has arisen in power generation scheduling recently, as the rapid increase in the number of gas-fired units has made power systems more vulnerable to failures in natural gas networks. The large-scale integration of wind power further exacerbates the problem because gas-fired units are usually scheduled to catch up wind power uncertainty and thus lead to great variations in the state of gas network. To meet this challenge, it is necessary to commit and dispatch the gas-fired units considering both wind uncertainty and natural gas network security. However, the dynamic characteristics of gas flow are remarkably slower than those of power flow, which should be appropriately modeled to explore its benefits for power system operation. Because directly applying partial differential equations overly complicates the already complicated generation scheduling problem, we address this problem by proposing an approximated transient matrix-form gas flow model. A two-stage robust generation scheduling model is then proposed considering the dynamic security constraints of gas networks and the wind power uncertainty. Moreover, we successfully avoid the nonlinearity of gas flow constraints by developing a new solution methodology. Finally, an illustrative case is presented to demonstrate the effect of gas network dynamics in generation scheduling.

Intelligent Parking Garage EV Charging Scheduling Considering Battery Charging Characteristic

Abstract: This paper studies the electric vehicle (EV) charging scheduling problem under a parking garage scenario, aiming to promote the total utility for the charging operator subject to the time-of-use (TOU) pricing. Different from most existing works, we develop a multicharging system incorporating the practical battery charging characteristic, and design an intelligent charging management mechanism to maximize the interests of both the customers and the charging operator. First, to ensure the quality of service for each client, we implement an admission control mechanism to guarantee all admitted EVs’ charging requirements being satisfied before their departure. Second, we formulate the charging scheduling process as a deadline constrained causal scheduling problem. Then, we propose an adaptive utility oriented scheduling (AUS) algorithm to optimize the total utility for the charging operator, which can robustly achieve low task declining probability and high profit. The charging operator can also apply the discussed reservation mechanism to mitigate the performance degradation caused by the charging information mismatching with vehicle stochastic arrivals. Finally, we conduct extensive simulations based on realistic EV charging parameters and TOU pricing. Simulation results exhibit the effectiveness of the proposed AUS algorithm in achieving desirable performance compared with other benchmark scheduling schemes.

Minimizing total carbon footprint and total late work criterion in flexible job shop scheduling by using an improved multi-objective genetic algorithm

Abstract: In scheduling, previous research attention has been directed towards classical-based objective functions, while ignoring environmental-based objective functions. The purpose of this research is to present a multi-objective flexible job shop scheduling problem with the objectives of minimizing total carbon footprint and total late work criterion, simultaneously, as sustainability-based and classical-based objective functions, respectively. In order to solve the presented problem effectively, an improved multi-objective genetic algorithm is proposed to obtain high quality non-dominated schedules. This work has three main scientific contributions that are: (1) This is a novel and pioneer research that addresses carbon footprint reduction in flexible job shop scheduling, (2) This is also the first research that addresses the total late work criterion in multi-objective flexible job shop scheduling, and (3) This research proposes an improved multi-objective evolutionary algorithm for solving the newly extended bi-objective problem. Stepwise delineation of the proposed algorithm is provided and fifteen newly extended test instances are solved by the proposed approach. Computational outcomes of the proposed algorithm are compared with two most representative and well-known multi-objective evolutionary algorithms, namely, non-dominated sorting genetic algorithm II and strength Pareto evolutionary algorithm 2. The principal results show that: (1) The proposed algorithm is superior in finding high quality non-dominated schedules, (2) It performs better in four averaged comparison metrics as compared to the other algorithms, and (3) Carbon footprint has an impact on the optimum solutions. As conclusions, the proposed algorithm is useful for production managers to schedule their operations in a way that can reduce carbon emission while minimizing late work. Production managers will also have the flexibility in selecting a schedule from amongst a set of non-dominated schedules.

A Self-Adaptive Sleep/Wake-Up Scheduling Approach for Wireless Sensor Networks

Abstract: Sleep/wake-up scheduling is one of the fundamental problems in wireless sensor networks, since the energy of sensor nodes is limited and they are usually unrechargeable. The purpose of sleep/wake-up scheduling is to save the energy of each node by keeping nodes in sleep mode as long as possible (without sacrificing packet delivery efficiency) and thereby maximizing their lifetime. In this paper, a self-adaptive sleep/wake-up scheduling approach is proposed. Unlike most existing studies that use the duty cycling technique, which incurs a tradeoff between packet delivery delay and energy saving, the proposed approach, which does not us duty cycling, avoids such a tradeoff. The proposed approach, based on the reinforcement learning technique, enables each node to autonomously decide its own operation mode (sleep, listen, or transmission) in each time slot in a decentralized manner. Simulation results demonstrate the good performance of the proposed approach in various circumstances.

A Collaborative Multiobjective Fruit Fly Optimization Algorithm for the Resource Constrained Unrelated Parallel Machine Green Scheduling Problem

Abstract: Due to the development of the green economy, green manufacturing has been a hot topic. This paper proposes a new problem, i.e., the resource constrained unrelated parallel machine green manufacturing scheduling problem (RCUPMGSP) with the criteria of minimizing the makespan and the total carbon emission. To solve the problem, a collaborative multiobjective fruit fly optimization algorithm (CMFOA) is proposed. First, a job-speed pair-based solution representation is presented, and an effective decoding method is designed. Second, a heuristic for initialization of the population is proposed. Third, three collaborative search operators are designed to handle three subproblems in the smell-based search phase, i.e., job-to-machine assignment, job sequence, and processing speed selection. The technique for order preference by similarity to an ideal solution and the fast nondominated sorting methods are both employed for multiobjective evaluation in the vision-based search phase. Moreover, a critical-path-based carbon saving technique is designed according to the problem analysis to further improve the nondominated solutions explored in the fruit fly optimization algorithm-based evolution. In addition, the effect of parameter setting is investigated and the suitable parameter values are recommended. Finally, numerical tests and comparisons are carried out using the randomly generated instances, which show that the CMFOA is able to obtain more and better nondominated solutions than other algorithms. The comparisons also demonstrate the effectiveness of the collaborative scheme and the carbon saving technique as well as the CMFOA in solving the RCUPMGSP.

Optimal Operation of Multimicrogrids via Cooperative Energy and Reserve Scheduling

Abstract: Microgrid (MG) represents one of the major drives of adopting Internet of Things for smart cities, as it effectively integrates various distributed energy resources. Indeed, MGs can be connected with each other and presented as a system of multimicrogrid (MMG). This paper proposes the optimal operation of MMGs by a cooperative energy and reserve scheduling model, in which energy and reserve can be cooperatively utilized among MMGs. In addition, values of Shapely are introduced to allocate economic benefits of the cooperative operation. Finally, a case study based on a system of MMGs is conducted, and simulation results verify the effectiveness of the proposed cooperative scheduling model.

IEEE 802.1Qbv Gate Control List Synthesis Using Array Theory Encoding

Abstract: Time Sensitive Networks (TSN) emerge as the set of sub-standards incorporating real-time support as an extension of standard Ethernet. In particular, IEEE 802.1Qbv defines a time-triggered communication paradigm with the addition of a time-aware shaper governing the selection of frames at the egress queues according to a predefined schedule, encoded in so-called Gate Control Lists (GCL). Nonetheless, the design of compositional systems with real-time demands requires a proper configuration of these mechanisms to truly achieve the temporal isolation of communication streams with end-to-end timeliness guarantees. In this paper we address how the synthesis of communication schedules for GCLs defined in IEEE 802.1Qbv can be formalized as a system of constraints expressed via first-order theory of arrays (T_A). We formulate the necessary constraints showing the suitability of the theory of arrays and discuss optimization opportunities arising from the underlying scheduling problem. Our evaluation using general-purpose SMT/OMT solvers proves the validity of the approach, scaling well for small-to medium-networks, and exposing trade-offs for the time needed to synthesize a schedule. Furthermore, we conduct a comparison against previous work and conclude the appropriateness of the method as the basis for future TSN scheduling tools.

Parallel bat algorithm for optimizing makespan in job shop scheduling problems

Abstract: Parallel processing plays an important role in efficient and effective computations of function optimization. In this paper, an optimization algorithm based on parallel versions of the bat algorithm (BA), random-key encoding scheme, communication strategy scheme and makespan scheme is proposed to solve the NP-hard job shop scheduling problem. The aim of the parallel BA with communication strategies is to correlate individuals in swarms and to share the computation load over few processors. Based on the original structure of the BA, the bat populations are split into several independent groups. In addition, the communication strategy provides the diversity-enhanced bats to speed up solutions. In the experiment, forty three instances of the benchmark in job shop scheduling data set with various sizes are used to test the behavior of the convergence, and accuracy of the proposed method. The results compared with the other methods in the literature show that the proposed scheme increases more the convergence and the accuracy than BA and particle swarm optimization.