Field workforce scheduling

BT employs thousands of field engineers across the UK to maintain networks, repair faults and provide service to customers. To allocate work more efficiently, BT launched Work Manager in the early 1990s — an information system automating work management and field communications, and now marketed as a.p.solve's TASKFORCE. In 1996, BT Exact's Intelligent Systems Lab enhanced Work Manager with a Dynamic Scheduler (DS) combining heuristic search and constraint-based reasoning. Since its national roll-out in 1997, DS has consistently reduced operational costs while preserving high quality of service. This paper gives an overview of BT's workforce scheduling problem, the DS system, and its operational and commercial impact. This revised version was published online in July 2006 with corrections to the Cover Date.     

Estimating Travel Times of Field Engineers

BT's Work Manager platform uses a dynamic scheduler to plan the daily jobs of field engineers. In order to produce reliable schedules, Work Manager requires accurate estimates about the time an engineer spends on travelling from one job location to the next (inter-job time) and how much time is required to complete a job. We have developed a generic platform called Intelligent Travel Time Estimation and Management System (ITEMS) that we have used to derive a specialised version for modelling the behaviour of BT's mobile workforce. This travel time estimation system (TTE) receives job data every night and learns a new estimation model on a daily basis. When a new estimation model significantly differs from the currently used model, the new estimation model is automatically uploaded into Work Manager. The estimation algorithm has been used in a loosely coupled trial for about two years. Currently, we are undertaking trials of TTE — a tightly coupled Web-based system that fully automates the learning and management of estimation models, providing a graphical user interface that displays rich detail about the travel patterns of BT's mobile workforce. During the trial period we have observed that using TTE can quickly improve the accuracy of travel time estimates by up to 10%. This revised version was published online in July 2006 with corrections to the Cover Date.     

Adaptive QoS scheduling in wireless cellular networks

Quality of service (QoS) has been always controversial in resource shared networks. Scheduling as a packet prioritizing mechanism at Data Link Layer (DLL) contributes to QoS guarantee provisioning significantly. In this paper, a novel packet scheduler is developed in wireless cellular networks. The proposed scheme provides QoS-guaranteed service for the applications running on the sensor nodes in all the three aspects of QoS, i.e. data rate, packet loss and packet delay with regard to jitter simultaneously. We establish a three-dimensional space with certain basis vectors for QoS and introduce the efficient point of performance in terms of QoS provisioning in that space. Then we develop a generalized metric, the QoS-deviation, which is the Euclidean distance between the QoS work point of flows and the QoS efficient point in the proposed space. Based on this metric, a novel scheduling approach, namely AQDC, is designed which makes it possible to tune the trade-off between QoS provisioning and throughput optimization in an adaptive manner depending on the current Cell QoS-deviation level (CDL). Furthermore, we also develop another scheduler, namely ARTC, which is the residual-time version of the AQDC scheduler. Finally, a QoS-deviation-based CAC policy will be introduced which can be applied to all schedulers without any consideration about their structure and can be employed in cellular packet switched networks.     

A novel scheduler for proportional delay differentiation by considering packet transmission time

The proportional delay differentiation model provides consistent packet delay differentiation between classes of service. The waiting time priority (WTP) scheduler is a priority scheduler in which the priority of a packet increases in proportion to its waiting time, and it is known as the best scheduler to achieve the proportional delay differentiation model. This paper proposes an advanced WTP (AWTP) scheduler, modified from WTP, that takes the packet transmission time into account. Simulation results reveal that when the link utilization is moderate (60%-90%), this scheduler not only obtains more accurate delay proportion than the WTP scheduler, but also reduces the average waiting time.     

ARMS — application of AI and OR methods to resource management

This paper documents work on automating resource management in BT Retail. BT Retail manages BT’s access network and provides telecommunications services to its customers. BT Retail’s field engineers are allocated jobs via an information system known as Work Manager. In order to proactively position the engineers (i.e. resources) so as to service jobs in an optimal manner, resource managers are involved in analysing the profiles of engineers in the light of incoming jobs and ’selecting’ those profiles that will yield best quality of service and reduce operational costs. A profile is a set of attributes that define a resource’s capabilities (i.e. skills), capacity (i.e. availability), and location (i.e. area). Resource planning involves identifying an ‘optimal’ set of resource profiles. Accurate workload forecasting is sine qua non for optimal resource planning. To this end we have developed ARMS, Automated Resource Management System, a suite of components for workload forecasting and resource planning.     

Space communications scheduler: A rule-based approach to adaptive deadline scheduling

Job scheduling is a deceptively complex subfield of computer science. The highly combinatorial nature of the problem, which is NP-complete in nearly all cases, requires a scheduling program to intelligently traverse an immense search tree to create the best possible schedule in a minimal amount of time. In addition, the program must continually make adjustments to the initial schedule when faced with last-minute user requests, cancellations, unexpected device failures, etc. A good scheduler must be quick, flexible, and efficient, even at the expense of generating slightly less-than-optimal schedules.

The Space Communications Scheduler (SCS) is an intelligent rule-based scheduling system developed at GE's Advanced Technology Laboratories. SCS is an adaptive deadline scheduler, which allocates modular communications resources to meet an ordered set of user-specified job requests on board the NASA Space Station. SCS uses pattern-matching techniques to detect potential conflicts within a schedule, then resolves these conflicts through algorithmic and heuristic means. As a result, the system generates and maintains high-density schedules without relying heavily on backtracking or blind search techniques. SCS was designed to allocate communication devices on board the Space Station, but its general-purpose scheduling strategy is suitable for many common real-world applications.     

Reliable real-time task scheduler based on Rocket Queue architecture

This paper presents the design of an improved task scheduler for real-time and safety-critical systems, where it is important to deal with real-time requirements and reliability requirements simultaneously. The proposed scheduler implements EDF algorithm for the optimal scheduling of hard real-time tasks, which is essential for real-time operating systems. The proposed task scheduler allows removing any task from the queue according to task ID and regardless of the actual position of the task within the queue, which is important for flexibility of the scheduler for its future extensions. Both operations of the scheduler, i.e. task adding and task killing take always constant time (two clock cycles) to execute regardless of the actual or the maximum number of tasks within the scheduler. The scheduler was verified using simplified version of UVM and applying millions of instructions with randomly generated sort values. The scheduler, implemented in a form of a coprocessor, was synthesized into Intel FPGA Cyclone V with 100 MHz clock frequency. There are two improvements proposed that can significantly reduce resource costs of the scheduler, which is achieved by replacing static deadlines with dynamic deadlines and using a new Rocket Queue architecture for sorting of the tasks according to their deadline values. When both improvements are applied simultaneously, the total ALM cost savings are in the range from 42,59% to 60,18% and the total amount of registers is reduced by 73,74% to 74,87%, depending on the scheduler capacity. The spared resources are then used for implementation of two different variations of TMR in order to increase fault tolerance of the scheduler. The resource cost reductions achieved also indirectly increase the reliability of such scheduler because of reduced probability that a fault occurs.     

Performance analysis of the adaptive EXP/PF channel scheduler in an AMC/TDM system

The adaptive EXP/PF AU; PL. DEFINE "EXP" ALSO channel scheduler proposed in Rhee et al. (2003) is an enhancement to the proportional fairness (PF) channel scheduler in that it guarantees delay sensitive services having the desired service delay outage probability as well as best-effort services in the forward link of an adaptive modulation and coding when used with time-division multiplexing (AMC/TDM) system. In this letter, assuming that there are many best-effort service users and one streaming service user requiring maximum delay constraints of 3 s, the system throughput of the adaptive EXP/PF scheduler is evaluated for different channel conditions of the streaming service user and is also compared with PF scheduler and its upper bound of throughput. The results show that the adaptive EXP/PF algorithm offers high system throughput as well as fairness among best-effort service users and guarantees streaming services of a desired delay constraint and outage probability.     

Opportunistic eigenbeamforming: Exploiting multiuser diversity and channel correlations

Multiuser diversity is an inherent form of diversity present in any time-varying system with several users. An opportunistic scheduler has to be used in order to exploit this type of diversity. A scheme that increases the effective dynamic range of the channel by deploying multiple antennas at the transmitter is called opportunistic beamforming. Opportunistic beamforming increases the degree of multiuser diversity in several scenarios, including correlated channels. Nevertheless, multiuser diversity can also be combined with other transmit schemes that have proven to be effective in correlated channels, such as eigenbeamforming. Eigenbeamforming is a point-to-point link transmit technique that could easily be combined with an opportunistic scheduler to extract multiuser diversity. We refer to the joint use of eigenbeamforming with an opportunistic scheduler as opportunistic eigenbeamforming. In this work, we show that the available multiuser diversity with opportunistic eigenbeamforming is larger than the one achieved when opportunistic beamforming is employed using the proportional fair scheduler under different degrees of correlation in the channel. In the present work we have considered a single-cell scenario.     

A heuristic approach to incremental and reactive scheduling

This paper describes a heuristic approach to incremental and reactive scheduling. Incremental scheduling is the process of modifying an existing schedule if the initial schedule does not meet its stated initial goals. Modifications made to a schedule during incremental scheduling typically consist of adding one or more activities by rescheduling existing activities. Reactive scheduling is performed when changes need to be made to an existing schedule due to uncertain or dynamic environments such as changes in available resources or the occurrence of targets of opportunity. Only minor changes are made during both incremental and reactive scheduling because a goal of rescheduling procedures is to minimally impact the schedule. A scheduling system generates a schedule in three phases. An initial batch scheduling phase, an incremental scheduling phase and a reactive scheduling phase. During the first phase, no rescheduling is attempted. All user requests are submitted to the scheduler and an initial schedule is created. During the second phase, non-computationally complex strategies must be used since the number of possible schedules that can be generated increases exponentially with the number of requests. Since simple strategies must be used for initial schedule creation, any schedule can potentially be greatly improved through the use of an incremental scheduling phase. Reactive scheduling occurs in near real-time in response to the occurrence of targets of opportunity. Consequently, a reactive scheduler must be able to generate schedules within acceptable time limits. Manual reactive scheduling is an inefficient strategy, and automated exhaustive search techniques are infeasible because of time limits. This paper describes the heuristic search techniques employed by the Request Oriented Scheduling Engine (ROSE), a prototype generic scheduler (Zoch & Hall, 1988). Specifically, we describe heuristics that efficiently approximate the cost of reaching a goal from a given state and effective mechanisms for controlling search.     

Optimal Burst Scheduling in Optical Burst Switched Networks

Optical burst switching (OBS) is an emerging technology that allows variable size data bursts to be transported directly over dense wavelength division multiplexing links. In order to make OBS a viable solution, the burst-scheduling algorithms need to be able to utilize the available wavelengths efficiently, while being able to operate fast enough to keep up with the burst incoming rate. For example, for a 16-port OBS router with 64 wavelengths per link, each operating at 10 Gb/s, we need to process one burst request every 78 ns in order to support an average burst length of 100 kB. When implemented in hardware, the well-known horizon scheduler has O(1) runtime for a practical number of wavelengths. Unfortunately, horizon scheduling cannot utilize the voids created by previously scheduled bursts, resulting in low bandwidth utilization. To date, minimum starting void is the fastest scheduling algorithm that can schedule wavelengths efficiently. However, while its complexity is O(log m), it requires 10 log m memory accesses to schedule a single burst. This means that it can take up to several microseconds for each burst request, which is still too slow to make it a practical solution for OBS deployment. In this paper, we propose an optimal burst scheduler using constant time burst resequencing (CTBR), which has O(1) runtime. The proposed CTBR scheduler is able to produce optimal burst schedules while having processing speed comparable to the horizon scheduler. The algorithm is well suited to high- performance hardware implementation.     

ARMS — Application of AI and OR Methods to Resource Management

This paper documents work on automating resource management in BT Retail. BT Retail manages BT's access network and provides telecommunications services to its customers. BT Retail's field engineers are allocated jobs via an information system known as Work Manager. In order to proactively position the engineers (i.e. resources) so as to service jobs in an optimal manner resource managers are involved in analysing the profiles of engineers in the light of incoming jobs and ‘selecting’ those profiles that will yield best quality of service (QoS) and reduce operational costs. A profile is a set of attributes that define a resource's capabilities (i.e. skills), capacity (i.e. availability), and location (i.e. area). Resource planning involves identifying an ‘optimal’ set of resource profiles. Accurate workload forecasting is sine qua non for optimal resource planning. To this end we have developed ARMS, Automated Resource Management System, a suite of components for workload forecasting and resource planning. This revised version was published online in July 2006 with corrections to the Cover Date.     

Intelligent Multi-modal Systems

Research in human/computer interaction has mainly focused on natural language, text, speech and vision primarily in isolation. Recently there have been a number of research projects that have concentrated on the integration of such modalities using intelligent reasoners. The rationale is that many inherent ambiguities in single modes of communication can be resolved if extra information is available.This paper describes an intelligent multi-modal system called the Smart Work Manager. The main characteristics of the Smart Work Manager are that it can process speech, text, face images, gaze information and simulated gestures using the mouse as input modalities, and its output is in the form of speech, text or graphics. The main components of the system are the reasoner, a speech system, a vision system, an integration platform and the application interface. The overall architecture of the system will be described together with the integration platform and the components of the system which include a non-intrusive neural network based gaze-tracking system. The paper concludes with a discussion on the applicability of such systems to intelligent human/computer interaction and lessons learnt in terms of reliability and efficiency.     

Improved MLWDF scheduler for LTE downlink transmission

In long-term evolution (LTE) downlink transmission, modified least weighted delay first (MLWDF) scheduler is a quality of service (QoS) aware scheduling scheme for real-time (RT) services. Nevertheless, MLWDF performs below optimal among the trade-off between strict delay and loss restraints of RT and non-RT traffic flows, respectively. This is further worsened with the implementation of hybrid automatic retransmission request (HARQ). As these restraints grow unabated with increasing number of user demands, the performance of MLWDF further reduces. In order to ameliorate this situation, there is a need to directly incorporate the variations in user demands and HARQ implementation as parameters to the MLWDF scheduler. In this work, an improvement to the MLWDF scheduler is proposed. The improvement entails adding two novel parameters that characterise user demand and HARQ implementation. The scheduler was tested using varying three classes of service in QoS class identifiers (QCIs) table standardised by Third Generation Partnership Project for LTE network to characterise different services. It was also tested on the basis of packet prioritisation. The proposed scheduler was simulated with LTE-SIM simulator and compared with the MLWDF and proportional fairness schedulers. In terms of delay, throughput and packet loss ratio; the proposed scheduler increased overall system performance.     

基于满意度准则采用混合波束成型预编码的调度方法

为了满足未来无线宽带网络中多种服务质量(QoS)要求业务并存的需求,该文提出一种多天线下行系统中结合混合波束成型预编码技术基于满意度的调度方法。该调度方法联合考虑用户的QoS要求及其信道质量,灵活地使用随机波束成型和迫零波束成型预编码技术,并通过惩罚机制避免特定用户占用超出其需求的带宽,更有效地为用户分配令其满意的资源,使各种QoS要求的用户获得满意的服务。仿真结果证明,与现有的方法相比,该文提出的方法在保证用户服务质量的同时,能更好地利用系统资源。     

An Approach to Scheduling Bursty Traffic

The scheduling scheme in packet switching networks is one of the most critical features that can affect the performance of the network. Hence, many scheduling algorithms have been suggested and some indices, such as fairness and latency, have been proposed for the comparison of their performances. While the nature of Internet traffic is bursty, traditional scheduling algorithms try to smooth the traffic and serve the users based on this smoothed traffic. As a result, the fairness index mainly considers this smoothed traffic and the service rate as the main parameter to differentiate among different sessions or flows. This work uses burstiness as a differentiating factor to evaluate scheduling algorithms proposed in the literature. To achieve this goal, a new index that evaluates the performance of a scheduler with bursty traffic is introduced. Additionally, this paper introduces a new scheduler that not only uses arrival rates but also considers burstiness parameters in its scheduling algorithms.     

Practical use of CAM in the IC industry

Based on an extensive improvement of cycle time, yield, defect density, equipment utilization and productivity, this paper will review in detail practical examples of data management and utilization through a CAM system. The basis for data utilization is made through the set of CAM reports which has to be designed for its usage by the manufacturers: shift changes, daily, weekly, monthly meetings; orientation towards problem solving; high level of commitment to organization goals. The short-cycle-time culture developed in our international IC manufacturing company has led us to enhance our WIP (Work in Process) management concept and our interaction with production control. It has therefore changed our requirements towards the CAM system especially in terms of scheduling: lot priority, WIP projection, start scheduler.     

JTMN: a Java-based TMN development and experimentationenvironment

Java technology has largely entered the world of element, network, and service management and has undoubtedly proven useful in combination with the SNMP framework for both manager and agent side developments. However, very few experiments have been made so far trying to provide an in-depth integration of this technology within the TMN framework. In this paper we present a free software which offers a generic environment which provides a basis for reaching this integration. This environment, called J^TMN, is composed of a set of Java software packages which enable both development and deployment of TMN compliant components. Not bound to any specific management platform and easily portable to any TMN system, its goal is to allow experimentation and deployment of more advanced features such as mobile agents, active network management, and advanced delegation in harmony with existing TMN approaches     

Large area MOS-gated power devices using fusible link technology

A new approach for improving the yield of large area MOS-gated power devices is described based upon wafer repair using fusible links of aluminum or polysilicon to isolate defective segments from the rest of the device. Unlike previously reported wafer repair techniques, the proposed approach does not require any knowledge of the location of the fault (gate-to-source short) within the device. Work done on the development of power-MOS process compatible fusible links is described in this paper. Power MOSFET's and IGBT's have been successfully fabricated using these fusible links to perform wafer repair     

Robust Packet Scheduling in Wireless Cellular Networks

This paper addresses the following robust scheduling problem: Given that only coarse-grained channel state information (i.e., bounds on channel errors, but not the fine-grained error pattern) is available, how to design a robust scheduler that ensures worst-case optimal performance? To solve this problem, we consider two coarse-grained channel error models and take a zero-sum game theoretic approach, in which the scheduler and the channel error act as non-cooperative adversaries in the scheduling process. Our results show that in the heavy channel error case, the optimal scheduler adopts a threshold form. It does not schedule a flow if the price (the flow is willing to pay) is too small, in order to maximize the system revenue. Among the scheduled flows, the scheduler schedules a flow with a probability inversely proportional to the flow price such that the risk of being caught by the channel error adversary is minimized. We also show that in the mild channel error model, the robust scheduling policy exhibits a balanced trade-off between a greedy decision and a conservative policy. The scheduler is likely to take a greedy decision if it evaluates the risk of encountering the channel error adversary now to be small. Therefore, robust scheduling does not always imply conservative decision. The scheduler is willing to take “risks” to expect higher gain in some scenarios. Our solution also shows that probabilistic scheduling may lead to higher worst-case performance compared to traditional deterministic policies. Finally, the current efforts show the feasibility to explore a probabilistic approach to cope with dynamic channel error conditions.