Edinburgh Napier University

  Proof of Work (PoW) is used to provide a consensus mechanism for Bitcoin. In this mechanism, the process of generating a new block in the blockchain is referred to as mining. Such process is intentionally designed to be resource-intensive and time consuming so that the rate of block generation remains steady. A single participant, called a miner usually has limited computation power to produce PoWs. This leads miners to form a mining pool, where miners aggregate their computing power and share the rewards. However, a phenomenon raises in such a mining pool activity, where miners attack each other. Consequently, this results in a decrease in total rewards received from the mining pool. To address the abovementioned problem, we build a multi-miner model for forming a mining pool. We further propose a method to improve the cooperation-probability of miners in the pool by introducing a Zero-Determinant strategy and a Temporal Difference learning method (TD(λ)). Experimental simulation results show that the proposed method can effectively promote the cooperation among miners, therefore, increase the rewards received from the formed mining pool.