Optimization statistical/mathematical equation to understand buying pattern and predict future revenue mix

A sales outlet selling 3 different products which are either bought individually or together, either as 2 or 3 have seen 7 different buying clusters (as shown in the attached 3 concentric rings). The outlet has been able to capture the pattern of sales from Jan to June for each of the 7 buying clusters. The business intent of the store is to ensure that more customers move into the innermost concentric ring of customers i.e. who buy the 3 products together. This is good because when customers buy all the 3 together, the likelihood of staying with the store is highest. This cluster of customers have been growing from January till date as the store offers special discounts for 2-in-1 or 3-in-1 purchases. The store now need to understand the pattern of change in the contribution of each of the 3 products and 7 buying patterns as seen from January to July to determine/develop the following: a) The relational logic and the optimal mix of the 7 buying clusters/3 products b) Optimization equation to determine future c) Predictive equation to determine optimal level of each of the 7 clusters for each month from September to December based on the historical pattern. The expected total revenue forecast for September –December is available buit need to understand how each element will optimally contribute if we develop an optimization equation using the 7 buying clusters.