Ho L Lee

7702561, Apr 20, 2010

Jul 31, 2001

09/917818

Andrew J. Davenport - Ossining NY, US
Jayant R. Kalagnanam - Tarrytown NY, US
Ho Soo Lee - Mount Kisco NY, US

International Business Machines Corporation - Armonk NY

G06Q 40/00

705 37, 705 7, 705 8, 705 28, 705 36 R

A method is provided for identifying a cost-minimizing bid set in a reverse combinatorial auction subject to various business rules for all-or-nothing bundled bids. The method includes introducing a decision variable for each bid, introducing a counting variable to indicate whether bids from a supplier are chosen in an optimal bid set, modeling demand constraints for each item using the bid variables, modeling minimum and maximum numbers of suppliers based on the counting variables, introducing dummy variables to ensure existence of feasible solutions, for a given cost formulating an objective of choosing bids that arrive early based on an additional timestamped objective with the given cost level as a constraint, and introducing price modifications to handle the formulated objective of choosing bids that arrive early.

20020065762, May 30, 2002

Mar 8, 2001

09/800664

Ho Lee - Mount Kisco NY, US
Juhnyoung Lee - Yorktown Heights NY, US

G06F017/60

705/037000, 705/026000

A method for purchasing and selling products or services in a networked environment using a request for quotation process and a visual interface for evaluating submitted bids for such products or services. A buyer submits a Request for Quotation (RFQ) and associated attributes and/or business rules over a network. As seller responds to the RFQ by submitting a bid with attribute values. A market maker uses the buyer attributes and/or business rules with the attribute values of the submitted bid to create a visual interface augmented by customized filters which are later used to evaluate seller submitted bids. The bids are received in the e-marketplace, at which time the e-marketplace can arrange, sort or filter the received bids in order to assist the buyer in examining and evaluating such bids. The filtering may include filtering an attribute value, an attribute line, a bid line or a portion of the bid line. Information may also be displayed associated with the bid line such as attribute information, general information or detailed information. The bid lines may also be tagged, counted, scrolled, enlarged or reduced.

49512250, Aug 21, 1990

Nov 14, 1988

7/272135

Ho S. Lee - Mount Kisco NY
Marshall I. Schor - Katonah NY

International Business Machines Corp. - Armonk NY

G06F 1518

364513

A pattern-matching system, such as of the RETE type, is updated. The updating operation is generalize to operate with correct match resuslts for generalized RETE networks (ones where join nodes may have either predecessor be another join node), and for arbitrary partial update state, characterized by stop nodes designating how far update operations have progressed for particular update tokens (designated suspended tokens) at any particular time. The updating operation consists of the following steps: First, all tokens only partially updated through the network are shadowed, in that the partial result memories associated with the stop nodes are put into the state that exists or existed before the suspended tokens arrive at the stop nodes. Then for each suspended token, it is pushed to resume nodes connected as RHS (or LHS) inputs in a descending order of resume nodes. The descending order is determined among sets of resume nodes where one node is reachable from the other, either in an upwards or downwards direction.

60443618, Mar 28, 2000

Mar 24, 1998

9/047275

Jayant R. Kalagnanam - Tarrytown NY
Ho Soo Lee - Mount Kisco NY
Mark Elliot Trumbo - Pleasantville NY

International Business Machines Corporation - Armonk NY

G06F 1700

705 28

A fast computer implemented method generates near-optimal solutions to the multi-objective inventory matching problem by solving for multiple objectives simultaneously and generating multiple non-dominating solutions. The method implements a multi-assignment backjumping algorithm that consists of three steps. The first step is a rappeling step in which a feasible solution is created by applying Iterative Bipartite Matching and maximum flow algorithm. Near-optimal feasible solutions are stored in a non dominated set. The second step is to use a multi-key sort to identify undesirable matches in a given feasible solution. The third step is backlifting the solution by removing undesirable matches from the feasible solution and places those undesirable matches on a no good set of matches. If the feasible solution is non-dominated, a copy is stored in a non-dominated set. The feasible solution is finally provided as input to the repelling step.

49567918, Sep 11, 1990

Nov 14, 1988

7/272097

Ho S. Lee - Mount Kisco NY
Marshall I. Schor - Katonah NY

International Business Machines Corp. - Armonk NY

G06F 1518

364513

A pattern-matching network is merged with an existing RETE network without recompiling the whole network, while still matching existing data. Permanent, temporary, full and partial merging operations are described, depending on the origination of match patterns among LHS, on demand (procedural programming), or top level. The merge process combines functionally identical nodes from the networks for reducing the number of nodes in the resulting network. The merger process produces so-called alpha and beta synapse nodes; alpha synapse nodes have pointers to top nodes, while beta synapse nodes consist pointers to a drain node, which is a node in the grafted RETE portions of the resulting network and to source memory nodes in the pre-existing RETE network. The algorithm for obtaining a set of synapse nodes without duplication of any logic connecting paths is described. The synapse nodes are used in updating a grafted network, also can be a suspended portion, upon completion of the merger.

58088911, Sep 15, 1998

Jul 30, 1996

8/688573

Ho Soo Lee - Mount Kisco NY
Mark Elliot Trumbo - Mount Kisco NY

International Business Machines Corporation - Armonk NY

G05B 1942

36446815

A scheduling method suitable for use in a primary steel production area operating in either direct rolling or direct hot charge rolling modes, collectively referred to as synchronized rolling. The steel production area comprising a continuous caster, for input of slabs to a hot strip mill. An order load for some finite period (e. g. one week) is received as input to the method. The method generates a caster schedule and a hot strip mill schedule which fills the order load with the objective of operating the continuous caster and the hot strip mill with minimal interruption. The generated schedule works within the constraints imposed by both the continuous caster and the hot strip mill. The schedule generation process also considers key operating constraints of individual facilities as well as those on which the facility has dependencies. The schedule also addresses objectives such as maximizing throughput, maximizing on-time delivery, and minimizing operating costs.

56664699, Sep 9, 1997

Feb 7, 1995

8/384979

Ho Soo Lee - Mount Kisco NY

International Business Machines Corporation - Armonk NY

G06F 1700

395 51

A method for generating solutions for sequencing problems of a type that can develop in artificial intelligence or operation research. The method includes the steps of finding multiple sets of subsequences of near optimal quality for a given finite number of n items wherein each item has a set of attributes and wherein there exists a set of constraints and objectives defined on the n items; and satisfying the constraints simultaneously such that the number of subsequences is a minimum.

63211327, Nov 20, 2001

Aug 5, 1998

9/129825

Milind Wasudeo Dawande - Peekskill NY
Jayant R. Kalagnanam - Tarrytown NY
Ho Soo Lee - Mount Kisco NY
Mark Elliot Trumbo - Pleasantville NY

International Business Machines Corporation - Armonk NY

G06F 1900

700 97

An efficient computer implemented method is used to design slabs for production from an order book. This method minimizes the number of slabs designed to fulfill an order book. This method is based on a heuristic algorithm which is a variant of the greedy approach for the set covering method. The variations are novel in three ways. First, designing slabs using the flexibility in the order size; second, using weight for choosing large slabs; and third, controlling the exponential nature of enumeration of the set of all subsets by constructing only the largest slab at each step.