US 12,271,949 B2
Linear model partitioner
Carl Erik Thornberg, Stockholm (SE)
Assigned to Chicago Mercantile Exchange Inc., Chicago, IL (US)
Filed by Chicago Mercantile Exchange Inc., Chicago, IL (US)
Filed on Jun. 18, 2024, as Appl. No. 18/746,772.
Application 18/746,772 is a continuation of application No. 17/404,400, filed on Aug. 17, 2021, granted, now 12,051,110.
Claims priority of provisional application 63/076,006, filed on Sep. 9, 2020.
Prior Publication US 2024/0338763 A1, Oct. 10, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 40/04 (2012.01); G05B 13/04 (2006.01); G06Q 40/06 (2012.01)
CPC G06Q 40/04 (2013.01) [G05B 13/041 (2013.01); G05B 13/042 (2013.01); G06Q 40/06 (2013.01)] 25 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
selecting, iteratively by a processor coupled with a memory, each of a plurality of different subsets of a plurality data records stored in a portfolio database in the memory, each data record characterized by a data size and storing data indicative of a position in one or more financial instruments, the selected subset being characterized by a magnitude correlated with the data size of the data records thereof, a value of each of one or more attributes, and one or more constraints thereon, each of the data records of the selected subset further including data indicative of an adjustable contribution by the position of the data record to the magnitude of the subset, data indicative of one or more contribution constraints on the adjustable contribution, and data indicative of an attribute contribution which specifies a degree to which the position of the data record contributes to the one or more attributes as a function of the adjustable contribution to the magnitude;
providing, by the processor, the selected subset, a specified goal and an approximation threshold to a linear optimizer coupled with the processor and operative to compute a solution comprising a modification to the magnitude of the selected subset which approximates the specified goal in accordance with the approximation threshold by determining one or more modifications to the adjustable contributions of one or more of the data records of the selected subset which are compliant with the contribution constraints thereon such that the value of each of the one or more attributes of the selected subset, as modified by the contribution thereto of each of the data records of the selected subset as a function of the modified adjustable contribution thereof, comply with the one or more constraints on the values of each of the one or more attributes, wherein a modified adjustable contribution of a particular data record of zero allows for removal of the particular data record from the portfolio database, reducing the data size of the portfolio database in accordance with the data size thereof; and
accumulating, by the processor in the memory for each selected subset, the received solutions from the linear optimizer until the accumulated solution satisfies a threshold, upon which applying, by the processor, the accumulated solution to the portfolio database to modify the data records stored therein in accordance therewith, wherein the data size of the portfolio database is reduced by removing those data records whose adjustable contribution of a particular data record was reduced to zero.