Can algorithms save energy?
An algorithm is a series of descriptions or instructions for executing a specific task or process. Among others, industry partners use algorithms devised at TUT in the development of their own applications. A good algorithm consumes as little power as possible.
We rely on sophisticated algorithms whenever we use mobile phones, tablets or other devices. Is it possible to create algorithms that reduce the power consumption of IT equipment?
Yes, it is. It takes less power for a device to run an energy-efficient algorithm than a conventional one. One way to achieve this is to optimize the algorithm, so that it requires as little processing power as possible.
Another option is to reduce the amount of data needed to execute an algorithm. For example, a device can determine a location based on various signals. If an algorithm is capable of producing positioning data without the continuous use of GPS, then closing the GPS receiver conserves power.
While it may be necessary to download data available online to run an algorithm, the process consumes power. It is possible to minimize the amount of data to be transmitted by tweaking the algorithm. The database may be stored on the device for future reference.
Where are energy-efficient algorithms used?
Energy-efficient algorithms are especially useful in portable, battery-powered devices. A well-designed algorithm extends their battery life and can help you get more juice out of other devices, too.
What must be taken into consideration when designing energy-efficient algorithms?
Algorithm designers must consider the processes that consume power and the expected level of precision. When there is a trade-off between speed and accuracy, they must come up with a reasonable compromise.
The question was answered by Dr. Matti Raitoharju, whose dissertation explored effective solutions for addressing challenging positioning problems. The methods he developed have wide applicability ranging from positioning and brain imaging to calculating the remaining range of an electric car.