High-speed multi-input tracker based on in-memory operations of time-dependent data

A non-transitory computer readable storage medium storing one or more programs for execution by one or more processors of an electronic device, the one or more programs, which, when executed by the one or more processors of the electronic device, cause the electronic device to:

• receive historical energy data for a plurality of groups of electrical devices;
• generate, for at least a subset of the historical energy data, a first logical table that corresponds to a first time, the first logical table for the first time including:
• (i) a plurality of logical columns, each logical column of the plurality of logical columns including an input vector of a plurality of input parameters corresponding to a respective time and (ii) a plurality of logical rows intersecting with the plurality of logical columns to define a plurality of logical cells, each logical row of the plurality of logical rows corresponding to, for the plurality of logical cells defined by the logical row, a respective input parameter of the plurality of input parameters, wherein each input parameter of the plurality of input parameters corresponds to an electrical parameter of the historical energy data;
• store the first logical table in a first memory;
• update logical cells in a respective logical column, of the first logical table for the first time stored in the first memory, with a first input vector that corresponds to a second time that is distinct from, and subsequent to, respective times corresponding to a plurality of logical columns in the first logical table stored in the first memory, thereby obtaining a second logical table that corresponds to the second time;
• obtain a first transposed kernel matrix for the second time between (i) the second logical table for the second time and (ii) the first input vector that corresponds to the second time;
• determine a first predicted output value for the second time based on a product of:
• (i) the first transposed kernel matrix for the second time, (ii) an inverse of a transposed kernel matrix for the first time, and (iii) a first mean vector for input vectors in the first logical table;
• and
• provide the first predicted output value for the second time to a power controller for setting electrical power parameters of one or more groups of electrical devices of the plurality of groups of electrical devices based on the first predicted output value.

A method for processing time-dependent data, comprising:

• at an electronic device that includes one or more processors and first memory storing a first logical table that corresponds to a first time, the first logical table for the first time including:
• a plurality of logical columns, each logical column of the plurality of logical columns including an input vector of a plurality of input parameters corresponding to a respective time;
• and
• a plurality of logical rows intersecting with the plurality of logical columns to define a plurality of logical cells, each logical row of the plurality of logical rows corresponding to, for the plurality of logical cells defined by the logical row, a respective input parameter of the plurality of input parameters:
• receiving historical energy data for a plurality of groups of electrical devices;
• generating, for at least a subset of the historical energy data, the first logical table that corresponds to the first time, wherein each input parameter of the plurality of input parameters corresponds to an electrical parameter of the historical energy data;
• updating logical cells in a respective logical column, of the first logical table for the first time stored in the first memory, with a first input vector that corresponds to a second time that is distinct from, and subsequent to, respective times corresponding to the plurality of logical columns in the first logical table stored in the first memory, thereby obtaining a second logical table that corresponds to the second time;
• obtaining a first transposed kernel matrix for the second time between (i) the second logical table for the second time and (ii) the first input vector that corresponds to the second time;
• determining a first predicted output value for the second time based on a product of:
• (i) the first transposed kernel matrix for the second time, (ii) an inverse of a transposed kernel matrix for the first time, and (iii) a first mean vector for input vectors in the first logical table;
• and
• providing the first predicted output value for the second time to a power controller for setting electrical power parameters of one or more groups of electrical devices of the plurality of groups of electrical devices based on the first predicted output value.