Skip to content

Latest commit

 

History

History
74 lines (52 loc) · 2.22 KB

README.md

File metadata and controls

74 lines (52 loc) · 2.22 KB
Raw

Tech Arena Baseline Description

Problem Statement

The problem is to manage a fleet of servers in a data center. The data center has four locations (DC1, DC2, DC3, DC4) and each location has a fleet of servers. The servers are of two types: CPU and GPU. Each server has a unique id, and a generation (S1, S2, S3, S4). The servers are managed in time steps. At each time step, the data center manager can take one of the following actions: buy, hold, move, and dismiss. The goal is to maximize the objective function of each data center by managing the servers efficiently.

See the problem statement pdf for more details.

Environment Setup

Install Conda by following the instructions here

For example, on Linux x86_64 (amd64), you can install Miniforge with the following command:

wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge-pypy3-Linux-x86_64.sh
sh Miniforge-pypy3-Linux-x86_64.sh
# And follow the instructions on the screen

Install the required packages by running the following commands:

cd starterkit/
mamba env create -f env.yml
conda activate fleet

Submission

To reproduce the baseline solution, run the following command:

cd experiments/buy_move_dismiss/
sh gen.sh

Evaluation

To evaluate your solution, you need to provide a json file with the following format:

{
    "fleet": [
        {
            "time_step": ,
            "datacenter_id": <DC1-DC4>,
            "server_generation": <CPU.S1-4 | GPU.S1-4>,
            "server_id": <server_id>,
            "action": <buy|hold|required|dissmiss>
        },
        ...
    ], 
    "pricing_strategy": {
        "time_step": ,
        "server_generation": <CPU.S1-4 | GPU.S1-4>,
        "latency_sensitive": <low|med|high>,
        "pricing": <pricing>
    }
}
// [{"time_step": 1, "datacenter_id": "DC1", "server_generation": "CPU.S1", "server_id": "7ee8a1de-b4b8-4fce-9bd6-19fdf8c1e409", "action": "buy"}]

The evaluation script is provided in the baseline/ directory. To evaluate your solution, run the following command:

fleet-eval -f <file_path> -s <seed>