Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: The invention provides for a computer-implemented method for solving a linear program (LP), the method comprising the steps of: receiving (100) the linear program; determining (101) a kernel K and determining a kernel matrix G of the kernel K, wherein the kernel matrix is a non-singular submatrix of the original matrix; determining (102) a set of non-basic variable indices and a set of extra constraint indices; computing (103) a primal kernel vector (xK) from the determined kernel; computing (104) a dual kernel vector (yK) from the determined kernel; and evaluating (105) the feasibility of the primal kernel vector and of the dual kernel vector.

Abstract: The invention provides for a computer-implemented method for solving a linear program (LP), the method comprising the steps of: receiving (100) the linear program; determining (101) a kernel K and determining a kernel matrix G of the kernel K, wherein the kernel matrix is a non-singular submatrix of the original matrix; determining (102) a set of non-basic variable indices and a set of extra constraint indices; computing (103) a primal kernel vector (xK) from the determined kernel; computing (104) a dual kernel vector (yK) from the determined kernel; and evaluating (105) the feasibility of the primal kernel vector and of the dual kernel vector.

Abstract: The invention provides for a computer-implemented method for solving a linear program (LP), the method comprising the steps of: receiving (100) the linear program; determining (101) a kernel K and determining a kernel matrix G of the kernel K, wherein the kernel matrix is a non-singular submatrix of the original matrix; determining (102) a set of non-basic variable indices and a set of extra constraint indices; computing (103) a primal kernel vector (xK) from the determined kernel; computing (104) a dual kernel vector (yK) from the determined kernel; and evaluating (105) the feasibility of the primal kernel vector and of the dual kernel vector.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: The invention provides for a computer-implemented method for solving a linear program (LP), the method comprising the steps of: receiving (100) the linear program; determining (101) a kernel K and determining a kernel matrix G of the kernel K, wherein the kernel matrix is a non-singular submatrix of the original matrix; determining (102) a set of non-basic variable indices and a set of extra constraint indices; computing (103) a primal kernel vector (xK) from the determined kernel; computing (104) a dual kernel vector (yK) from the determined kernel; and evaluating (105) the feasibility of the primal kernel vector and of the dual kernel vector.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: A method for deterministic locking in a parallel computing environment is provided. The method includes creating a data structure in memory of a computer for a shared resource. The data structure encapsulates a reference to an owner of a lock for the shared resource and a queue of threads able to seek exclusive access to the shared resource. The queue in turn includes different entries, each entry including an identifier for a corresponding one of the threads and a deterministic time computed for the corresponding one of the threads from a count of memory accesses occurring in the corresponding one of the threads. Consequently, a thread can be selected from the queue to receive ownership of the lock and exclusive access to the shared resource based upon a deterministic time for the selected thread as compared to other deterministic times for others of the threads in the queue, for example, a lowest deterministic time.

Abstract: An optimization engine includes a mixed-integer programming (MIP) solver that receives a programming model, an outcome objective, and a group of start vectors. Each of the MIP start vectors in the group specify one or more restrictions to apply to the programming model. The MIP solver uses the programming model to compute a potential solution from each of the MIP start vectors included in the group, which results in a group of potential solutions. Next, the MIP solver selects one of the potential solutions in the group as an optimal intra-group solution. The optimal intra-group solution is the potential solution in the group that best achieves the outcome objective. In turn, the optimal intra-group solution is used to complete the outcome objective.

Abstract: The present invention relates to methods and systems for generating solutions to global optimization problems. In one aspect, the invention allows for determining whether models of optimization problems are infeasible. In another aspect, the invention allows for identifying relatively small sets of constraints that make a model infeasible. In yet another aspect, the invention provides methods and systems for creating one or more artificial infeasibilities in a model so as to seek improvement upon a known value of the objective function. Creation of artificial constraints in this manner may also permit identification of a relatively small set of constraints that may be limiting the value of the objective function.

Abstract: An optimization engine includes a mixed-integer programming (MIP) solver that receives a programming model, an outcome objective, and a group of start vectors. Each of the MIP start vectors in the group specify one or more restrictions to apply to the programming model. The MIP solver uses the programming model to compute a potential solution from each of the MIP start vectors included in the group, which results in a group of potential solutions. Next, the MIP solver selects one of the potential solutions in the group as an optimal intra-group solution. The optimal intra-group solution is the potential solution in the group that best achieves the outcome objective. In turn, the optimal intra-group solution is used to complete the outcome objective.

Abstract: The present invention relates to methods and systems for generating solutions to global optimization problems. In one aspect, the invention allows for determining whether models of optimization problems are infeasible. In another aspect, the invention allows for identifying relatively small sets of constraints that make a model infeasible. In yet another aspect, the invention provides methods and systems for creating one or more artificial infeasibilities in a model so as to seek improvement upon a known value of the objective function. Creation of artificial constraints in this manner may also permit identification of a relatively small set of constraints that may be limiting the value of the objective function.