DEER
From PRAGMAgridWIKI
(Please don't change in this page. Click edit, then cut/paste the wiki contents to a new page for your application, and edit them in your new page.)
Contents |
Introduction
The West Nile Encephalitis WNE started in New York in northeast USA in 1990 but quickly spread to the southwest in a matter of 3-4 years (Figure 1). The local scale of mosquito dispersion cannot account for this speed with which the disease spread. The spread was indeed enhanced by infected birds that may fly over large distances of tens of km over a season or a few months. This pattern of dispersion may be conceptualized in Figure 2, in which uneven patches of distributions (squares) spreading over a rectangle of 3000 km by 4000 km are used to represent WNE distribution on the scale of continental USA. Over a local patch (a square) of the order of a few km, the PDE for mosquito will be solved by finite difference method with mesh size of the order of 0.1 km, each patch requiring a total number of mesh of the order of 10 to 100 thousands nodes. In this setup, grid technology is appropriate, where a grid will simulate each local habitat (lined square).
For large system simulations, normal computing power offered by the regular PC would not be adequate. We therefore propose the application of high power computing such as parallel or grid computing to provide the needed power. We hope that this project would stimulate active research and collaboration in this region.
 Figure 1 Distribution of WNE |  Figure 2 Conceptual computation grid |
DEER Description
The Dengue and Encephalitis Eradication Routines DEER is a numerical simulation model designed to simulate the dynamics of vector borne disease transmission. First, we focus on the distribution of Aedes aegypti mosquito population and the dynamics of dengue disease transmission. Dengue disease has only one epidemiological cycle linking the human hosts and the vector mosquitoes.
DEER model is based on a mass balance for two phases of mosquito lifecycle: the winged female mosquito (represented by W) and the aquatic form, which includes eggs, larvae and pupae (represented by A). For disease transmission, the winged female mosquitoes (W) divided into two different classes, susceptible wing (WS) and infected wing (WI). Humans are another state variables and it has 4 different sub-variables, susceptible human (HS), infected human (HI), recovered human (HR) and dead human (HD). The compartment model for Dengue disease transmission is in Figure 3. Humans get dengue virus infections from the bite of an infected wing mosquito. Wing mosquitoes become infected when they bite infected humans.
Then, we further extended DEER model to simulate distributions of other vector borne diseases such as WNE and Japanese Encephalitis (JE). WNE disease transmission involves another mammal like birds or horse in the epidemiological cycle linking human host and mosquito. Bird is a new state variable for DEER. Similar with human, it has 4 sub-variables such as susceptible birds (BS), infected birds (BI), recovered birds (BR) and dead birds (BD). Figure 4 showed that West Nile Virus Disease has a complex life cycle that mainly circulates between mosquitoes and birds. The virus is transmitted to humans by a mosquito bites. Humans are a dead end host in the life cycle of the encephalitis virus.
The algorithm in DEER has been coded by MPI commands that some routines can be calculated paralelly. We have successfully run DEER. We have tried to run some cases using our USM Campus-Grid which has 22 numbers of CPU on the cluster.
Application driver and contact
1. Ms. Tan Kah Bee 2. Dr. Teh Su Yean PhD student Lecturer School of Mathematical Sciences School of Mathematical Sciences Universiti Sains Malaysia Universiti Sains Malaysia 11800 Penang, Malaysia 11800 Penang, Malaysia avery701040@gmail.com su_yean@hotmail.com; syteh@usm.my
Applicaiton Requirements
Hardware requirements
- At least 10 GB Disk storage for data storage
- At least 25 number of CPU
Software requirements
- GCC Compiler
o The source codes of DEER use Fortran language. It requires the GCC Compiler $ gcc -v Reading specs from /usr/lib/gcc/i386-redhat-linux/3.4.6/specs Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable -__cxa_atexit --disable-libunwind-exceptions --enable-java-awt=gtk --host=i386-redhat-linux Thread model: posix gcc version 3.4.6 20060404 (Red Hat 3.4.6-3)
- MPI library
o MPI is required for parallel execution of DEER.
Site access status
Application Run
Execution Script
lamboot machine /opt/mpich/gnu/bin/mpif77 deer.for -o deer /opt/mpich/gnu/bin/mpirun -np 10 -machinefile machine deer
Simple Demonstration
Small Case
The simple distribution of Aedes aegypti mosquito population has been showed at Figure 5. The research areas are a rectangle of 40 km by 40 km. We assumed that small concentration of wing mosquito have been introduced at central of research areas on day 0. The wing mosquitoes spread out if the location has water container which is suitable for mosquito lay eggs. Small numbers of mosquitoes are frequently carried by vehicles to other locations. It spread out, if the location is suitable for mosquitoes. We assumed that people will spray insecticide to kill the mosquitoes at certain time. However the density of insecticide decrease with time after it sprayed. When insecticide is sprayed, the distribution of mosquito decreases temporarily, but eventually it returns to the pretreatment level.
Figure 6 shows the physical simulation time when various numbers of CPUs are used. The time it takes to run a test case becomes saturated after 12 numbers of CPU. This is because the execution speed of the program is limited by the system’s I/O latencies such as file access, message passing over network. Figure 7 shows the memory storage needed for different cases. The memory storage required increase exponentially as the calculation domain increases.
Large Case
We hope that we can use the PRAGMA grid to run the simulation for the dynamics of disease transmission of WNE disease which has propagated over whole USA in 3-4 years. The research areas are a big rectangle of almost 3000 km by 4000 km. However, the flight distance of mosquito could range only from a few hundred meters to less than 25 m in an urban area. The proper mesh size of DEER is in order to less than 1 km. We normally use 0.1 km. Because of the large research areas and high memory storage required for the WNE simulation, it is better using high power computing to reduce the time cost.
Publication
1. Tan, K.B., Koh, H.L. and Teh, S.Y. (2009). Modeling Transmission of Dengue Fever .To be presented at the 6th International Conference on Fuzzy Systems and Knowledge Discovery, 14-16 August 2009, Tianjin, China.
2. Sui, L.L., Teh, S.Y., Koh, H.L. and Izani, A.M.I. (2009). Stage-Structured Cohort Model for Mosquitoes.To appear in the Proceedings of 5th Asian Mathematical Conference (AMC), 22-26 June 2009, PWTC, Kuala Lumpur, Malaysia.
3. Tan, K.B., Teh, S.Y., Koh, H.L. and Izani, A.M.I. (2009). Dengue Fever and Climate Change. To appear in the Proceedings of 5th Asian Mathematical Conference (AMC), 22-26 June 2009, PWTC, Kuala Lumpur, Malaysia.
4. Sui, L.L., Koh, H.L., Tan, K.B., Teh, S.Y. and Izani, A.M.I. (2009). Modeling Dengue Infection for Malaysia. Presented at 16th Pacific-Rim Application And Grid Middleware Assembly (PRAGMA16), 23-26 March, Daejeon, Korea.
5. Tan, K. B., Teh, S. Y., Koh, H. L., Sui, L. L., Bahari, B. and Izani, A.M.I. (2009) Modeling West Nile Virus with Grid Technology. Presented at 16th Pacific-Rim Application And Grid Middleware Assembly (PRAGMA 16), 23-26 March, Daejeon, Korea.
6. Koh, H. L., Tan, K. B., Sui, L. L., Teh, S. Y., Bahari, B. and Izani, A.M.I. (2008). Modeling Mosquitoes Distribution. Presented at 15th Pacific-Rim Application And Grid Middleware Assembly (PRAGMA 15), 22-23 October, Penang, Malaysia.
7. Koh, H.L., Teh, S.Y., Tan, K.B. and Izani, A.M.I. (2008). Simulation of Mosquito Distributions under Climate Change. Proceedings of the 6th Regional IMTGT UNINET Conference: ‘Sustaining Natural Resources Towards Enhancing the Quality of Life Within the IMT-GT Zone’, 28-30 August 2008, Penang, Malaysia, p. 339-343.
8. Teh, S.Y., Koh, H.L., Tan, K.B. and Izani, A.M.I. (2008). COHORT Model for Simulating Mosquito Distributions. Proceedings of the 6th Regional IMTGT UNINET Conference: ‘Sustaining Natural Resources Towards Enhancing the Quality of Life Within the IMT-GT Zone’, 28-30 August 2008, Penang, Malaysia, p. 352-356.
9. Koh, H.L., Teh, S.Y., Izani, A.M.I. and DeAngelis, D.L. (2008). Modeling Biological Invasion: The Case of Dengue and Mangrove. Invited Lecture in International Conference on Mathematical Biology – ICMB07. American Institute of Physics Conference Proceedings, Volume 971, New York, p. 11-18.
10. Tan, K.B., Koh, H.L., Izani, A.M.I. and Teh, S.Y. (2008). Modeling Mosquito Population With Temperature Effects. Proceedings of the International Conference on Environmental Research and Technology (ICERT’ 08), 28-30 May 2008, Penang, Malaysia, (Eds) Teng Tjoon Tow, Fera Fizani Ahmad, Norli Ismail, Fazilah Ariffin and Anees Ahmad, Penerbit Universiti Sains Malaysia, p. 536-540.
11. Koh, H.L., Lee, H.L., Teh, S.Y. and Izani, A.M.I. (2007). Environmental and Ecological Modeling: Perspectives and Prospects. Keynote Speech. To appear in the Proceedings of 2nd Regional Conference on Ecological and Environmental Modeling (ECOMOD 2007), 28-30 August, Penang, Malaysia.
12. Koh, H.L., Teh, S.Y. and Izani, A.M.I. (2007). DEER Dengue and Encephalitis Eradication Routines: Some Observations. To appear in the Proceedings of the 2nd Regional Conference on Ecological and Environmental Modeling (ECOMOD 2007), 28-30 August, Penang, Malaysia.
13. Tan, K.B., Koh, H.L., Izani, A.M.I. and Teh, S.Y. (2007). Modeling Mosquito Population Subject to Environmental Capacity. Proceedings of the 3rd IMT-GT Regional Conference on Mathematics, Statistics and Applications, 5-6 December 2007, Universiti Sains Malaysia, p. 1085-1091.
14. Koh, H.L., Lee, H.L., Teh, S.Y., Saw, S.K. and Izani, A.M.I. (2006). Biomathematics: From Daphnia to Dengue. Proceedings of the National Conference on Mathematical Biology (KKMB06), 22-23 August 2006, invited by Universiti Putra Malaysia (UPM), Institute for Mathematical Research (INSPEM), 8 p.
