| 1. |
- Behravesh, Rasoul, et al.
(författare)
-
ML-Driven DASH Content Pre-Fetching in MEC-Enabled Mobile Networks
- 2020
-
Ingår i: 16th International Conference on Network and Service Management, CNSM 2020, 2nd International Workshop on Analytics for Service and Application Management, AnServApp 2020 and 1st International Workshop on the Future Evolution of Internet Protocols, IPFuture 2020. - : Institute of Electrical and Electronics Engineers Inc.. - 9783903176317
-
Konferensbidrag (refereegranskat)abstract
- Streaming high-quality video over dynamic radio networks is challenging. Dynamic adaptive streaming over HTTP (DASH) is a standard for delivering video in segments, and adapting its quality to adjust to a changing and limited network bandwidth. We present a machine learning-based predictive pre-fetching and caching approach for DASH video streaming, implemented at the multi-access edge computing server. We use ensemble methods for machine learning (ML) based segment request prediction and an integer linear programming (ILP) technique for pre-fetching decisions. Our approach reduces video segment access delay with a cache-hit ratio of 60% and alleviates transport network load by reducing the backhaul link utilization by 69%. We validate the ML model and the pre-fetching algorithm, and present the trade-offs involved in pre-fetching and caching for resource-constrained scenarios.
|
|
| 2. |
|
|
| 3. |
- Bohlin, Markus, et al.
(författare)
-
A Tool for Gas Turbine Maintenance Scheduling
- 2009. - 20
-
Ingår i: Proceedings of the Twenty-First Conference on Innovative Applications of Artificial Intelligence (IAAI'09). - : IEEE Computer Society. - 9781577354239
-
Konferensbidrag (refereegranskat)abstract
- We describe the implementation and deployment of a software decision support tool for the maintenance planning of gas turbines. The tool is used to plan the maintenance for turbines manufactured and maintained by Siemens Industrial Turbomachinery AB (SIT AB) with the goal to reduce the direct maintenance costs and the often very costly production losses during maintenance downtime. The optimization problem is formally defined, and we argue that feasibility in it is NP-complete. We outline a heuristic algorithm that can quickly solve the problem for practical purposes, and validate the approach on a real-world scenario based on an oil production facility. We also compare the performance of our algorithm with results from using mixed integer linear programming, and discuss the deployment of the application. The experimental results indicate that downtime reductions up to 65% can be achieved, compared to traditional preventive maintenance. In addition, using our tool is expected to improve availability with up to 1% and reduce the number of planned maintenance days with 12%. Compared to a mixed integer programming approach, our algorithm not optimal, but is orders of magnitude faster and produces results which are useful in practice. Our test results and SIT AB’s estimates based on operational use both indicate that significant savings can be achieved by using our software tool, compared to maintenance plans with fixed intervals.
|
|
| 4. |
|
|
| 5. |
- Bohlin, Markus, et al.
(författare)
-
Searching for gas turbine maintenance schedules
- 2010
-
Ingår i: The AI Magazine. - : Wiley. - 0738-4602. ; 31:1, s. 21-36
-
Tidskriftsartikel (refereegranskat)abstract
- Preventive-maintenance schedules occurring in industry are often suboptimal with regard to maintenance coallocation, loss-of-production costs, and availability. We describe the implementation and deployment of a software decision support, tool for the maintenance planning of gas turbines, with the goal of reducing the direct maintenance costs and the often costly production losses during maintenance down time. The optimization problem is formally defined, and we argue that the feasibility version is NP-complete. We outline a heuristic algorithm that can quickly solve the problem for practical purposes and validate the approach on a real-world scenario based on an oil production facility. We also compare the performance of our algorithm with result's from using integer programming and d'iscuss the deployment of the application. The experimental results indicate that down time reductions up to 65 percent can be achieved, compared to traditional preventive maintenance. In addition, the use of our tool is expected to improve availability by up to 1 percent and to reduce the number of planned maintenance days by 12 percent. Compared to an integer programming approach, our algorithm is not optimal but is much faster and produces results that are useful in practice. Our test results and SIT AB's estimates based on operational use both indicate that significant savings can be achieved by using our software tool, compared to maintenance plans with fixed intervals.
|
|
| 6. |
- Gillblad, Daniel, et al.
(författare)
-
Estimating the Parameters of Randomly Interleaved Markov Models
- 2009
-
Konferensbidrag (refereegranskat)abstract
- Sequences that can be assumed to have been generated by a number of Markov models, whose outputs are randomly interleaved but where the actual sources are hidden, occur in a number of practical situations where data is captured as an unlabeled stream of events. We present a practical method for estimating model parameters on large data sets under the assumption that all sources are identical. Results on representative examples are presented, together with a discussion on the accuracy and performance of the proposed estimation algorithms. Finally, we describe a real-world case study where we apply the technique to the sequence of events recorded in the technical support database of an IT vendor.
|
|
| 7. |
|
|
| 8. |
- Gillblad, Daniel, et al.
(författare)
-
Fault-Tolerant Incremental Diagnosis with Limited Historical Data
- 2008
-
Konferensbidrag (refereegranskat)abstract
- We describe a novel incremental diagnostic system based on a statistical model that is trained from empirical data. The system guides the user by calculating what additional information would be most helpful for the diagnosis. We show that our diagnostic system can produce satisfactory classification rates, using only small amounts of available background information, such that the need of collecting vast quantities of initial training data is reduced. Further, we show that incorporation of inconsistency-checking mechanisms in our diagnostic system reduces the number of incorrect diagnoses caused by erroneous input.
|
|
| 9. |
- Gillblad, Daniel, et al.
(författare)
-
Fault-tolerant incremental diagnosis with limited historical data
- 2006. - 1
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- In many diagnosis situations it is desirable to perform a classification in an iterative and interactive manner. All relevant information may not be available initially and must be acquired manually or at a cost. The matter is often complicated by very limited amounts of knowledge and examples when a new system to be diagnosed is initially brought into use. Here, we will describe how to create an incremental classification system based on a statistical model that is trained from empirical data, and show how the limited available background information can still be used initially for a functioning diagnosis system.
|
|
| 10. |
- Gonzales Prieto, Alberto, et al.
(författare)
-
Toward Decentralized Probabilistic Management
- 2011
-
Ingår i: IEEE Communications Magazine. - 0163-6804 .- 1558-1896. ; 49:7, s. 80-96
-
Tidskriftsartikel (refereegranskat)abstract
- In recent years, data communication networks have grown to immense size and have been diversified by the mobile revolution. Existing management solutions are based on a centralized deterministic paradigm, which is appropriate for networks of moderate size operating in relatively stable conditions. However, it is becoming increasingly apparent that these management solutions are not able to cope with the large dynamic networks that are emerging. In this article, we argue that the adoption of a decentralized and probabilistic paradigm for network management will be crucial to meet the challenges of future networks, such as efficient resource usage, scalability, robustness, and adaptability. We discuss the potential of decentralized probabilistic management and its impact on management operations, and illustrate the paradigm by three example solutions for real-time monitoring and anomaly detection.
|
|
