I'm trying to understand the paper Incentive Compatibility of Bitcoin Mining Pool Reward Functions (Schrijvers, Bonneau, Doneh and Roughgarden, in Financial Cryptography and Data Security – FC 2016 Workshops, BITCOIN, 2016; PDF).
In page 3 Section 2.1, they say pool operator does not know actual $\alpha_i$ mining power of player $i$. To estimate $\alpha_i$ depends on the reported shares and solutions.
A reward function $ R\colon H \mapsto [0,1]^n$ is a function from a history transcript to an allocation $ \{a_i\}_{i=0} ^ {n} $ with $ \sum_i a_i =1 $. I don't understand this? What allocation does the authors mean?
What I somewhat understand from next para is $ H(k) = b = (y_1(k), \cdots, y_i(k) ) $ where $y_i(k)$ is no of shares reported by player $i$ in round $k$. This I am basing on $H$ contains for each miner $i$ the total no of shares $b _i$ reported in that round. History transcript is given by a vector $b \in N^n $.
Also what do they mean when they say " We use vector notation for $b$, so $b_1 + b_2 $ means component wise addition of these, and $\|b\|_1 = \sum_{i=1}^n b_i $ is the sum of components of $b$"
Could someone explain with examples or in a more concrete way ? Also any suggestions for understanding this paper would be much appreciated. Thanks.