Power Trading Framework of Cloud Edge Computing in an Artificial Intelligence Bazaar

Mekala S.; Ravuri A.; Kumar K.S.; Gupta S. · Advances in AI for Cloud, Edge, and Mobile Computing Applications (2025)

Resumen

In the present world, technologies are spurred by the rapid expansion of Internet of Things and the flourishing progress in deep learning technology; we are witnessing a continuous proliferation of applications of Artificial Intelligence. Simultaneously, advancements in established computing paradigms, namely edge computing and cloud computing, offers range of computing solution to meet the ever-growing demands for widespread AI services. However, several complex issues persist within these computing frameworks that include the inefficient utilization of computing resources, self-interested behavior in power trading computing mechanisms as well as inefficiency in management of AI services. To address these challenges, we introduce a blockchain-based framework for trading computing power, which we refer to as AI-Bazaar. Multiple roles are assumed with AI consumers in AI bazaar and have the flexibility for both services of AI and mining using blockchain technology. Consequently, the study formulates the problem of trading computing resources as a Stackelberg game. Lever aging the WoLF principle (Win or Learn), we have devised a PB-MARL algorithm to seek equilibrium within AI-Bazaar, ensuring equitable profits for both AI consumers and CPPs. We have conducted numerical simulations to illustrate the effectiveness of the proposed framework.