LOW-LATENCY APPROXIMATION OF COMBINATORIAL OPTIMIZATION OF RESIDUAL AMOUNTS WHEN ALLOCATING LARGE COLLECTIONS OF STORED VALUE CARDS

1 . A method, comprising:

obtaining, with one or more computers, a repository of stored value card records, each record corresponding to a stored value card, and each record including an identifier of a merchant accepting the respective stored value card and a balance of value remaining on the respective stored value card, each stored value card having at least one previous user possessor;

receiving, with one or more computers, a request for a stored value card from a remote client computing device;

inferring, with one or more computers, based on the request, a transaction balance of a transaction in which the requested stored value card is to be used;

inferring, with one or more computers, based on the request, a merchant to participate in the transaction;

identifying, with one or more computers, a subset of the stored value card records in response to determining that the subset of stored value cards are accepted by the inferred merchant, the subset having more than 20 stored value cards with more than 15 different balances;

selecting, with one or more computers, one or more stored value cards from among the identified subset based on a comparison between the inferred transaction balance and the value remaining on the selected one or more stored value cards; and

sending, with one or more computers, balance-access information by which the selected one or more stored value card balance or balances can be spent at a point of sale terminal.

2 . The method of claim 1 , wherein:

the repository stores records of more than 10,000 stored value cards, subsets of which being accepted by more than 10 merchants geographically distributed over a country;

the stored value cards include gift cards that were transferred from an earlier possessor to a gift card exchange;

some of the stored value cards have balance-access information accessed by more than two previous user possessors;

more than 10 requests for stored value cards are received per minute from user devices among more than 500,000 user computing devices requesting stored value cards over a one year duration;

the comparison includes comparing combinations of four or more stored value cards; and

sending the balance-access information is performed within ten seconds of receiving the request for the stored value card;

3 . The method of claim 1 , wherein inferring a transaction balance comprises:

inferring a transaction balance based on previous transactions with the inferred merchant.

4 . The method of claim 1 , wherein inferring a transaction balance comprises:

inferring a transaction balance based on previous transactions by a user requesting the stored value card.

5 . The method of claim 1 , wherein inferring a transaction balance comprises inferring a transaction balance based on:

previous transactions by a user requesting the stored value card; and

previous transactions with the inferred merchant.

6 . The method of claim 1 , wherein inferring a transaction balance comprises performing steps for inferring a transaction balance.

7 . The method of claim 1 , wherein inferring a merchant comprises:

obtaining a geolocation of a mobile computing device associated with the request; and

ascertaining the merchant based on the geolocation.

8 . The method of claim 1 , wherein inferring a merchant comprises performing steps for inferring the merchant.

9 . The method of claim 1 , wherein:

selecting one or more stored value cards is performed according to a greedy selection algorithm that accounts for an amount of times at least some of the stored value cards are possessed by users.

10 . The method of claim 1 , wherein:

after receiving the request, selecting one or more stored value cards comprises selecting a sub-optimal set of one or more stored value cards to provide one or more stored value cards more optimal than these selected in response to a different request.

11 . The method of claim 1 , wherein selecting one or more stored value cards comprises:

selecting a stored value card among the subset in response to determining that the selected stored value card has a balance closest to a threshold balance based on the inferred transaction balance among the identified subset of stored value cards.

12 . The method of claim 1 , wherein selecting one or more stored value cards comprises:

selecting plurality of stored value card among the subset in response to determining that a cumulative balance of the selected plurality of stored value cards have a balance closest to a threshold balance based on the inferred transaction balance.

13 . The method of claim 1 , wherein selecting one or more stored value cards comprises:

determining balances associated with more than 500 candidate responses, each of the candidate responses being associated with one or more of the identified subset of the stored value cards, some of the candidate responses being associated with two of the identified subset of the stored value cards, each of the candidate responses having a cumulative balance associated with the respective stored value card or cards associated therewith; and

selecting among the candidate responses based on differences between the inferred transaction balance and the balances of the candidate responses.

14 . The method of claim 13 , wherein selecting among the candidate responses based on differences between the inferred transaction balance and the balances of the candidate responses comprises:

selecting among the candidate responses based on an amount of previous possessors of balance-access information of at least some of the stored value cards associated with the respective candidate responses.

15 . The method of claim 13 , wherein the candidate responses comprise:

each combination of two stored value cards among the identified subset of stored value cards.

16 . The method of claim 13 , wherein the candidate responses comprise:

each permutation of two stored value cards among the identified subset of stored value cards.

17 . The method of claim 13 , wherein the candidate responses comprise:

each combination of four stored value cards among the identified subset of stored value cards, and wherein the identified subset of stored value cards comprise more than 1,000 stored value cards, and wherein balance-access information is sent within ten seconds of receiving the request.

18 . The method of claim 13 , wherein determining balances associated with more than 500 candidate responses is performed in advance of receiving the request.

19 . The method of claim 1 , wherein selecting one or more stored value cards comprises performing steps for selecting one or more stored value cards based on the transaction balance amount.

20 . A system, comprising:

one or more processors; and

memory storing instructions that when executed by at least some of the processors effectuate operations comprising:

obtaining a repository of stored value card records, each record corresponding to a stored value card, and each record including an identifier of a merchant accepting the respective stored value card and a balance of value remaining on the respective stored value card, each stored value card having at least one previous user possessor;

receiving a request for a stored value card from a remote client computing device;

inferring based on the request, a transaction balance of a transaction in which the requested stored value card is to be used;

inferring based on the request, a merchant to participate in the transaction;

identifying a subset of the stored value card records in response to determining that the subset of stored value cards are accepted by the inferred merchant, the subset having more than 20 stored value cards with more than 15 different balances;

selecting one or more stored value cards from among the identified subset based on a comparison between the inferred transaction balance and the value remaining on the selected one or more stored value cards; and

sending balance-access information by which the selected one or more stored value card balance or balances can be spent at a point of sale terminal.