China
Africa
Explore chapters and articles related to this topic
Artificial Intelligence for Network Operations
Published in Mazin Gilbert, Artificial Intelligence for Autonomous Networks, 2018
Artificial intelligence permeates the change management process—from optimization algorithms used to calculate and recalculate schedules to the execution and validation of the effectiveness of change activities. One particularly challenging area that AI is addressing is the change scheduling. Maintenance activities that are potentially customer impacting are typically performed during the maintenance window—typically night hours when traffic is low and thus customer impact is low. This gives a limited window in which activities can be scheduled; with complex network topologies, multiple different maintenance activities making changes to network functions and with the large network scale, such scheduling can be complex. Customer impact must be minimized—thus coordinating across different network functions of different types and across different network layers is crucial to ensuring that redundant service paths are not simultaneously impacted. As higher priority activities arise, or as maintenance activities go awry or network impairments interfere with maintenance activities, schedules need to be reworked, adding to the complexity. Such scheduling is a complex coordination problem to determine across change activities which times are available for activities on a given network function. Algorithms to schedule change activities are implemented in ONAP on top of the optimization framework—these algorithms take into account service paths and other constraints to ensure that change activities being simultaneously executed on different network functions do not cause unnecessary customer impact.
Optimal opportunistic tamping scheduling for railway track geometry
Published in Structure and Infrastructure Engineering, 2021
Hamid Khajehei, Mohammad Haddadzade, Alireza Ahmadi, Iman Soleimanmeigouni, Arne Nissen
The unused life of track section in the given maintenance window is calculated using this formula: where is the unused life of track section in maintenance window and are the degradation rate and SDLL value for track section in maintenance window respectively, and is the upper bound of the (corrective) maintenance limit.
Mixed Biogeography-Based Optimization for GENCOs’ Maintenance Scheduling in Restructured Power Systems
Published in Applied Artificial Intelligence, 2018
Abdolvahhab Fetanat, Gholamreza Shafipour
Constraints Equation (10) represent the maintenance window stated in terms of the start of maintenance variables . The unit and the line must be available both before their earliest period of maintenance and their latest period of maintenance, e.g., . Set of constraints Equation (11) consists of crew and resource availability, seasonal limitation and desirable schedule. The seasonal limitation can be incorporated into and values of constraint Equation (10). If we consider for example that lines 1, 2 and 3 are to be maintained simultaneously, the set of constraints can be formed as follows:
Predictive maintenance scheduling with reliability characteristics depending on the phase of the machine life cycle
Published in Engineering Optimization, 2021
Iwona Paprocka, Wojciech M. Kempa, Bożena Skołud
An uncertainty model for which information on a disruption is described using numerical ranges is presented by Bali and Labdelaoui (2015). The time horizon of a schedule is divided into periods in which maintenance tasks are carried out continuously. The optimal allocation of maintenance that guarantees a high level of reliability and reduces both production and maintenance costs is searched. The perturbation mechanism alters a start time of maintenance randomly within the allowed maintenance window. Bajestani, Banjevic, and Beck (2014) consider two cases: the state (production rate) of the machine is known at the beginning of a time period; or the probability that the machine is in a given state depends on both the machine condition and the time of performing maintenance. The average machine production rate is the average of the expected production rate of the machine at the beginning and at the end of the period. The average values are described by given time intervals (uniformly distributed) with no historical analyses. Mokhtari, Mozdgir, and Abadi (2012) develop an integrated production and maintenance scheduling model for a parallel machine system. They consider different kinds of preventive maintenance activities corresponding to the partial service or full repair of each machine. The availability of the machine is deteriorated with the maximum repair rate described by the exponential distribution. Maintenance intervals are randomly inserted into the schedule using two kinds of move: maintenance policy swapping and maintenance policy perturbation. Lei (2011) investigates a fuzzy job-shop scheduling problem with an availability constraint for maintenance. Some maintenance operations occur on each machine during a planning horizon. The author does not describe how the number of maintenance operations is estimated. Moreover, each maintenance window has a fixed predefined start time and duration.