Predictive Scheduling Research Articles

Machine learning algorithms benchmarking for real-time fault predictable scheduling on a shop floor

To select a proper machine learning algorithm for fault predictable scheduling on a shop floor, ten algorithms in the machine learning field have been selected, implemented and compared in this research. Due to the lack of applicable real data to the authors, a data generation method is proposed in terms of data complexity, number of data attributes and data depth. On top of the method, six datasets are generated by selecting three-level data attributes and three-level data depths, which were used to train the ten algorithms. The performances of the algorithms are evaluated by considering three indexes including, training accuracy, testing time and training time. The results demonstrate that naive Bayes classifier is suitable to low-complexity data and that convolutional neural network and deep belief network fit well in high-complexity data, such as the real data. [Submitted 25 June 2018; Accepted 27 May 2019]

Availability-based predictive maintenance scheduling for vibrating-grate biomass boilers

This paper proposes an availability-based maintenance scheduling for a vibrating-grate biomass boiler. The main objective is to minimise the maintenance programme cost while maintaining the system availability at a targeted level. A modified failure density function was employed to deliberate the effect of maintenance on the reliability of the components. The constant failure rate of the components which yields constant hazard functions were converted to time-dependent hazard functions by means of Weibull regression analysis. The updated failure parameters were examined by Kolmogorov Smirnov tests and the results showed a good agreement with the primary data. A critical factor named maintenance impact plays an important role in optimised scheduling and determines which components have a higher influence on system availability. Fault tree analysis method was engaged to estimate system availability. The results showed 32% reduction in the number of maintenance requirements with the help of modified predictive maintenance management.

Model-based multi-objective predictive scheduling and real-time optimal control of energy systems in zero/low energy buildings using a game theory approach

Online optimal control of energy systems plays a significant role in achieving the zero/low energy goal and high energy efficiency for zero/low energy buildings. Increasing amount of studies are conducted on the multi-objective optimal control of building energy systems in recent years since multiple objectives are often concerned in practice. The weighted sum method is widely used to solve the multi-objective optimization problems for online optimal controls. However, it is often difficult and even impractical to determine proper weights for objectives of different natures. Existing studies on online optimal control of energy systems in zero/low energy buildings focused on the predictive scheduling, and very few studies addressed the real-time optimal control aspect which is also essential in operation. In this study, a coordinated online multi-objective control strategy, consisting of two control optimization schemes, is therefore proposed for predictive scheduling and real-time optimal control of energy systems in zero/low energy buildings. A cooperative game theory-based method is adopted for the online multi-objective optimizations. The control strategy was tested and evaluated via simulation of the energy systems in a zero energy building with battery storage on two typical days. Control variables and weights of objectives were optimized to minimize the combined optimization objective involving energy cost and grid impact. The test results show that it is essential and beneficial to coordinate the predictive scheduling and real-time optimal control in actual operation. The cooperative game theory-based method is effective for the online multi-objective optimization without the need of setting weights of different objectives.

Predictive maintenance scheduling with reliability characteristics depending on the phase of the machine life cycle

The method of maintenance and production task scheduling is proposed to make the production system insensitive to a disruption. The method enables the maintenance of high values of schedule evaluation indicators in variable conditions, such as stability robustness and quality robustness. The frequency of a machine failure is described by the Maxwell distribution. Formulae for different reliability characteristics are given. The parameters of the distribution are estimated using maximum likelihood, empirical moments and a method based on the renewal theory. The advantage of the last method is the ability to achieve accurate predictions of failure-free time of a machine, when there is a lack of complementary and reliable historical data. In the related literature, only the first method is used to estimate unknown parameters of the probability distribution. The methods are applied to job-shop scheduling problems to evaluate their impact on achieving predictive schedules.

Using Xen and KVM as real-time hypervisors

The recent developments in virtualisation technologies have made feasible the execution of complex and performance-critical applications in virtual machines. Some of such applications are characterised by real-time constraints and require a predictable scheduling of virtual machines on physical cores, hence several works in real-time literature have proposed advanced scheduling and design techniques to respect the application constraints. This paper complements those works, investigating the latencies introduced by two of the most widely used open-source hypervisors, Xen and KVM. Some guidelines for properly configuring the VMs in order to reduce the introduced latencies (so that previous theoretical analysis and algorithms can be used in practice) are also provided, showing that both KVM and Xen are usable as real-time hypervisors.

Robust predictive–reactive scheduling and its effect on machine disturbance mitigation

Matrix-structured production systems are becoming increasingly appealing for realizing an individualized mass production. An appropriate method for scheduling in such systems under consideration of machine disturbances is necessary due to the increased complexity and degrees of freedom in order to achieve a high and stable performance. A robust predictive–reactive method for scheduling is proposed and the relationship between increased robustness in predictive scheduling and the ability to reschedule is analyzed in this paper. The result will enhance understanding of the utility of robust scheduling methods in order to maintain high logistic performance even when facing machine disturbances.

Development of Smart Rural Village Indicators in Line With Industry 4.0

In 1950, approximately 1.8 billion lived in rural areas and subsequently, rural population growth averaged 1.0 per cent per year, with the result that the rural population nearly doubled, reaching 3.4 billion in 2018 and is expected to peak in 2021 at just over 3.4 billion. Rural villages are lacking behind urban areas when it come to the stages of industrial revolutions and societal development with respect to technological advancement. Today, the world is faced with the new wave of advanced technological revolution called Industry 4.0. Despite the impressive work being implemented successfully to create smart cities and smart villages in various countries, there are still no developed standard indicators that can be used to define smart rural village concept. The objective of the study is to develop indicators for smart rural villages aligned with Industry 4.0 technologies using systematic literature review, the 2030 agenda for sustainable development, as well as the ISO 37122, smart cities indicator's standard. The translation method developed prioritized indicators according to sustainability, smartness and connectivity from the Information and Communication Technology. The main contribution is the method and indicators developed over 2-year period for smart rural villages. The resulting indicators makes it possible to answer what a smart rural village is: village that has access to affordable energy, adequate housing, safe drinking water, an inquisitive rural culture, early warning systems against adverse weather conditions, against drinking water pollution as well as the predictable schedule for local doctor, taxi/bus.

Online prediction of tray-transport request time using mechanistic grey box models for improved scheduling of robotic strawberry harvest-aids

This paper addresses the online prediction of the time when a strawberry picker will fill up the carton tray (s)he is currently harvesting with. This time is important for predictive scheduling of tray-transporting harvest-aid robots that transport trays instead of the pickers themselves, with the goal to increase efficiency. To make such predictions, two mechanistic grey models (GM1 and GM2) were developed. The models incorporate knowledge about the process of filling a tray with harvested fruits by estimating relevant process parameters. During harvesting, both models compute Bayesian updates of the a priori distributions of the picking time and the tray's mass when it is empty and full, with data from each full tray. Moreover, both models rely on real-time measurements of the mass of harvested crop to predict when the tray capacity will be reached. GM1 predicts the next tray transport request (NTTR) time by adding the a priori mean picking time obtained from a previously developed predictive linear mixed model (LMM), and the beginning time of a cycle, whereas, GM2 fits a simple linear regression (SLR) to the load cell data in real time. In order to evaluate the performances of the two models, a set of load cell output time series was obtained from an instrumented picking cart during manual strawberry harvesting. Except for one tray (out of 58), both GM1 and GM2 were successful at generating NTTR predictions. However, GM2 predictions adapted to the occurrence of non-picking activities that manifested as plateaus in the mass data. Also, when a tray was not filled completely, GM1 could not predict accurately the NTTR time of that tray neither the NTTR of the next tray, because it could not detect the starting time for filling the next tray. GM2 missed the starting time also, but was able to provide accurate predictions, albeit a bit later in the picking cycle. However, delays that do not extend beyond 70% of the picking time are not significant. For these reasons, it seems that GM2 should be preferred over GM1. Additionally, since both GM algorithms utilise the a priori statistics of picking time and the full tray's mass, it is advisable to have pickers deliver a couple of trays themselves in the beginning of their shift, so that the process parameters are “learned” with Bayesian updating.

Formulation and Comparison of Two Real-Time Predictive Gear Shift Algorithms for Connected/Automated Heavy-Duty Vehicles

This paper examines the problem of predictive gear scheduling for fuel consumption minimization in connected/automated heavy trucks. The literature highlights the fuel economy benefits of such predictive scheduling, but there is a need to optimize such scheduling online, in real time. To address this need, we begin by using dynamic programming (DP) to schedule gear shifting offline, in a manner that achieves a globally optimal Pareto tradeoff between the conflicting objectives of minimizing fuel consumption and shift frequency. The computational cost of DP is unfavorable for online implementation, but we present two algorithms addressing this challenge. Both algorithms rely on the fact that in the Pareto limit where fuel consumption minimization is the sole objective, DP furnishes a simple static shift map . Our first algorithm trains a recurrent neural network to prune the shift schedule generated by this map. The second algorithm performs this pruning in a direct manner tailored to reduce the schedule's rain flow count. We simulate these algorithms for different drive cycles. Both algorithms achieve a reasonable tradeoff between fuel consumption and gear shift frequency. However, the rain flow count algorithm is both more effective in approaching the DP-based Pareto front and more computationally efficient.

CoMM: a consensus algorithm for multi-agent-based manufacturing system to deal with perturbation

The emergence of Cyber Physical System has dramatically impacted the use of traditionally centralized control system in responding to unexpected events. Rush order is a quite common unexpected event in the current dynamic market characteristics and has significant perturbing ability to a centrally predictive schedule. This paper is aimed to propose a consensus algorithm for multi-agent-based manufacturing system (CoMM) to control the rush order and henceforth minimize a makespan. Consensus is an algorithmic procedure applied in control theory which allows convergence of state between locally autonomous agents collaborating for their common goal. Leader-follower communication approach was used among the multi-agent to deal with the perturbing event. Each agent decides when to broadcast its state to neighbor agents, and the controlling decision depends on the behavior of this state. The consensus algorithm is initially modeled by networking all contributing agents. After this, it is validated with simulation experiment based on academic full-sized ap plication platform called TRACILOGIS platform. The results showed that the consensus algorithm has significantly minimized the impact of rush order on makespan of manufacturing orders launched on a system.

Resource-Oriented Partitioning for Multiprocessor Systems with Shared Resources

Predictable scheduling and resource sharing primitives are fundamental aspects of real-time systems. To prevent race conditions, access to shared resources must ensure mutual exclusion, e.g., using semaphores. Further, real-time locking protocols are required to avoid un-controlled priority inversions. For uniprocessor systems, the Priority Ceiling Protocol (PCP) has been widely accepted and supported in real-time operating systems. However, it remains arguable as to whether there exists a preferable approach for resource sharing in multiprocessor systems. In this paper, we show that the proposed Resource-Oriented Partitioned (ROP) scheduling with a distributed resource sharing policy, originating from the concept of the Distributed Priority Ceiling Protocol (DPCP), can achieve a non-trivial speedup factor guarantee. Specifically, we prove that the proposed R-PCP-rm-rm algorithm achieves a speedup factor of $11-6/(m+1)$11-6/(m+1) on a platform consisting of $m$m processors, where each job of a task may request at most one shared resource at most one time. Our empirical evaluations show that the proposed algorithm is highly effective in terms of task sets deemed schedulable.

Slippery Fish: Enforcing Regulation Under Subversive Adaptation

Attempts to curb illegal activity through regulation gets complicated when agents can adapt to circumvent enforcement. Economic theory suggests that conducting audits on a predictable schedule, and (counter-intuitively) at high frequency, can undermine the effectiveness of audits. We conduct a large-scale randomized controlled trial to test these ideas by auditing Chilean vendors selling illegal fish. Vendors circumvent penalties through hidden sales and other means, which we track using mystery shoppers. Instituting monitoring visits on an unpredictable schedule is more effective at reducing illegal sales. High frequency monitoring to prevent displacement across weekdays to other markets backfires, because targeted agents learn faster and cheat more effectively. Sophisticated policy design is therefore crucial for determining the sustained, longer-term effects of enforcement. A simpler demand-side information campaign generates two-thirds of the gains compared to the most effective monitoring scheme, it is easier for the government to implement, and is almost as cost-effective. The government subsequently chose to scale up that simpler strategy.

Coupling of centralized and decentralized scheduling for robust production in agile production systems

Individualized products and timely delivery require agile just-in-time manufacturing operations. Scheduling needs to deliver a robust performance with high and stable results even when facing disruptions such as machine failures. Existing approaches often generate predictive schedules and adjust them reactively as disturbances occur. However, the effectiveness of rescheduling approaches highly depends on the available degrees of freedom in the predictive schedule. In the proposed approach, a centralized robust scheduling procedure is coupled with a decentralized reinforcement learning algorithm in order to adjust the required degrees of freedom for a maximally efficient production control in real-time.

Prediction of variable technological operation times in production jobs scheduling

This paper presents a methodology for determining the variability of technological operations times, which is suitable for predictive scheduling of production tasks. The paper begins with outlining the problems related to production scheduling in real production systems and the process of robust job scheduling. In the work, a special attention was paid to the uncertainties associated with the variability of the times of technological operations of individual jobs. Subsequently, the proprietary prediction algorithm for the variability of technological operations times was discussed, and the results of its operation were verified using historical production data. The work was completed with the analysis of the obtained results, as well as the outlining the direction of further work in the analysed problem area.

Robust and stable flexible job shop scheduling with random machine breakdowns: multi-objectives genetic algorithm approach

In this paper, robust and stable scheduling for a flexible job-shop problem with random machine breakdowns has been discussed. A two-stage genetic algorithm is used to generate the predictive schedule. The first stage optimises the primary objective, which minimises the makespan, where all data is considered to be deterministic with no expected disruptions. The second stage optimises two objectives, makespan and stability, function in the presence of random machine breakdowns. For the second stage two different versions of multi-objective genetic algorithm, non-dominated sorting genetic algorithm II and non-dominated ranking genetic algorithm, is used. A simulator is proposed to simulate random machine breakdowns. An experimental study and analysis of variance is conducted to study the results of each multi-objective algorithm and breakdown simulator. The results of their comparison indicate that, non-dominated ranking genetic algorithm (NRGA) performs better and also shows a significant difference between various repair times in the proposed breakdown simulator.

Diszkrét termelési folyamatok újraütemezési feladatainak megoldása többcélú kereső algoritmussal

The paper summarizes the features of rescheduling problems of the discrete production processes and describes a new extended solving method. The purpose of the rescheduling is to manage the unexpected events occurred during the manufacturing. The rescheduling process creates a modified version of the predictive schedule for a given time horizon. It is required that the performance indicators take up suitable values and the deviation from the original schedule should be as small as possible. To solve this problem, we propose an advanced multi-objective searching method by using locking techniques.

A Cross-Layer SCTP Scheme With Redundant Detection for Real-Time Transmissions in IEEE 802.11 Wireless Networks

In this paper, we propose a Cross-Layer Stream Control Transmission Protocol (CL-SCTP), which supports multi-streaming and partial reliability for real-time transmissions in IEEE 802.11 wireless networks. In the application layer, we design a Screening Overdue MEssages (SOME) algorithm to leave out transmissions of probable overdue messages. An overdue message refers to a message exceeding its delay budget before the message can be transmitted in the transport layer. Then in the MAC layer, we propose the MOnitoring REdundant frames (MORE) algorithm to detect and remove probable redundant frames. A retransmitted frame from the upper layer makes the same frame entering the MAC layer earlier a redundant frame. We discuss our scheme in detail and perform simulations for both random losses and burst losses. Simulation results show our scheme performs better than Cross-Layer Forwarding Prediction Scheduling (CL-FPS), partially reliable SCTP (PR-SCTP), and SCTP in terms of efficient delivery ratio, average goodput, average end-to-end delay, and representative video metrics including Peak Signal to Noise Ratio (PSNR), Structural Similarity (SSIM), and Video Quality Metric (VQM).

Timely-Throughput Optimal Scheduling With Prediction

Motivated by the increasing importance of providing delay-guaranteed services in general computing and communication systems, and the recent wide adoption of learning and prediction in network control, in this paper, we consider a general stochastic single-server multi-user system and investigate the fundamental benefit of predictive scheduling in improving timely-throughput, being the rate of packets that are delivered to destinations before their deadlines. By adopting an error rate based prediction model, we first derive a Markov decision process (MDP) solution to optimize the timely-throughput objective subject to an average resource consumption constraint. Based on a packet-level decomposition of the MDP, we explicitly characterize the optimal scheduling policy and rigorously quantify the timely-throughput improvement due to predictive-service, which scales as Θ(p[C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> (a - a <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">max</sub> q)ρ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">τ</sup> /(p -q)+C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> (1-(1/p)](1-ρ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">D</sup> )), where a, a <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">max</sub> , ρ ∈ (0, 1), C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> > 0, C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ≥ 0 are constants, p is the true-positive rate in prediction, q is the false-negative rate, r is the packet deadline, and D is the prediction window size. We also conduct extensive simulations to validate our theoretical findings. Our results provide novel insights into how prediction and system parameters impact performance and provide useful guidelines for designing predictive low-latency control algorithms.

Two-stage teaching-learning-based optimization method for flexible job-shop scheduling under machine breakdown

In the real-world situations, uncertain events commonly occur and cause disruption of normal scheduled activities. Consideration of uncertain events during the scheduling process helps the organizations to make strategies for handling the uncertainties in an effective manner. Therefore, in the present paper, unexpected machine breakdowns have been considered during scheduling of jobs in a flexible job-shop environment. The objective is to obtain lowest possible makespan such that robust and stable schedules are produced even if an unexpected machine breakdown occurs. The robust and stable schedules may help to decrease the costs associated with unexpected machine failures. The present work uses a two-stage teaching-learning-based optimization (2S-TLBO) method to solve flexible job-shop scheduling problem (FJSP) under machine breakdown. In the first stage, the primary objective of makespan is optimized without considering any machine breakdown. In the second stage, a bi-objective function considering robustness and stability of the schedule is optimized under uncertainty of machine breakdowns. In order to incorporate the machine breakdown data to basic FJSP, a non-idle time insertion technique is used. In order to generate effective robust and stable predictive FJSP schedules, a rescheduling technique called modified affected operations rescheduling (mAOR) is used. The Kacem’s and Brandimarte’s benchmark problems have been solved and compared with other algorithms available in the literature. Results indicate that TLBO outperforms other algorithms by generating superior robust and stable predictive schedules. Statistical analysis is carried out to test the significance difference of the results obtained by TLBO with other algorithms.

A Predictive Map Task Scheduler for Optimizing Data Locality in MapReduce Clusters

This article describes how data locality is becoming one of the most critical factors to affect performance of MapReduce clusters because of network bisection bandwidth becomes a bottleneck. Task scheduler assigns the most appropriate map tasks to nodes. If map tasks are scheduled to nodes without input data, these tasks will issue remote I/O operations to copy the data to local nodes that decrease execution time of map tasks. In that case, prefetching mechanism can be useful to preload the needed input data before tasks is launching. Therefore, the key challenge is how this article can accurately predict the execution time of map tasks to be able to use data prefetching effectively without any data access delay. In this article, it is proposed that a Predictive Map Task Scheduler assigns the most suitable map tasks to nodes ahead of time. Following this, a linear regression model is used for prediction and data locality based algorithm for tasks scheduling. The experimental results show that the method can greatly improve both data locality and execution time of map tasks.