Abstract: A system that avoids dead-end states during a real-time heuristic search. While transitioning from a previous state to a current state, the system may perform lookahead to populate a state-space with potential states. The system may identify safe states using a safety predicate and only select potential states that are ancestors of safe states, providing a clear path to safety when needed. In addition, the system may determine a distance-to-safety function that indicates a number of state transitions between each potential state and a nearest safe state.

Abstract: Systems and methods are described that facilitate performing model-based planning techniques for allocations of multi-capacity resources in a machine. The machine may be, for instance, a printing platform, such as a xerographic machine. According to various features, the multi-capacity resource may be a sheet buffer, and temporal constraints may be utilized to determine whether an insertion point for a new allocation of the sheet buffer is feasible. Multiple insertion points may be evaluated (e.g., serially or in parallel) to facilitate determining an optimal solution for a print job or the like.

Abstract: A printing system includes a plurality of print media processing modules which transfer print media therebetween during printing and a fault management agent associated with each of the modules for acquiring fault-related data from the respective processing module. A fault management system is in communication with the fault management agents and receives fault-related data from the fault management agents. The fault management system processes the fault related data to identify faults in the system, such as when a first of the processing modules is a cause of fault-related data acquired in a second of the processing modules. When a fault is identified, a reconfiguration agent may reconfigure the printing system to mitigate an impact of at least one of the identified faults.

Abstract: A computer-based method and system for AI planning based quasi-Monte Carlo simulation for probabilistic planning are provided. The method includes generating a set of possible actions for an initial state, generating a set of sample future outcomes, generating solutions for each of the sample future outcomes, using an AI planner, generating a set of future outcome solutions that are low probability and high-impact, combining the solutions generated from each of the sample future outcomes with the future outcome solutions generated by the AI Planner into an aggregated set of future outcome solutions, analyzing the aggregated set of future outcome solutions, selecting a best action based at least partially on the analysis of the aggregated set of future outcome solutions, and outputting the selected best action to computer memory.

Abstract: Features described herein relate to concurrently processing multiple batches of job requests for one or more machines and/or components thereof, using a plurality of job planning queues. Each batch of job requests is allocated to a planning queue, and each planning queue comprises an unplanned subqueue that stores unplanned jobs, an unsent subqueue that stores planned jobs waiting to be executed, and a sent subqueue that stores planned jobs that have been output to the machine(s) for execution. A job planner and related components determine which unsent subqueue has the fewest planned jobs at a given point in time, and selects an unplanned job from the unplanned subqueue in the same planning queue as the identified unsent subqueue. The planner then generates a plan for the selected job and inserts the planned job into the unsent subqueue for eventual output to the machine(s) for execution.

Abstract: A method and tool is provided to obtain an optimistic estimate or exact optimal value of an operational parameter for a realistic system model under investigation. The model includes components and paths arranged to process continuous or discrete commodities. The system could be a model of a manufacturing system with different machines processing multiple job types, with different sequences of operations at different stages. Constraints are applied to the abstracted network flow model, and a plurality of steady state network flows are performed. and combined to captures the transformation of the commodities from a first state to a final output state. The optimistic estimate of the realistic system model under investigation is then returned through use of a general purpose or custom solver. The method can be used to perform tradeoff studies between machine allocations, job mixes, operating costs, reliability and throughput, or to speed up scheduling and machine control.

Abstract: A system that re-plans jobs based at least in part on user preferences in response to system component errors includes an exception handler that receives an exception from one of a plurality of components executing a plan to process a job. The system further includes a planner that creates a new plan for the job based at least in part on a model-based planning technique and at least one user preference.

Abstract: A combinatorial search method and system is implemented in a computer control system for utilizing state-space planning of operations for multi-step production processes. The planner considers various possible combinations of actions, searching for one that correctly transforms the initial state of the object (or commodity) into the specified desired final state, where each combination of actions the planner considers is called a search node. Each node contains a plan representing a series of actions of a plurality of machines on a single object and also containing the predicted state of the object with those actions applied either forward or backward. The state of the object consists of the set of attributes of the object.

Abstract: A computer-based method and system for AI planning based quasi-Monte Carlo simulation for probabilistic planning are provided. The method includes generating a set of possible actions for an initial state, generating a set of sample future outcomes, generating solutions for each of the sample future outcomes, using an AI planner, generating a set of future outcome solutions that are low probability and high-impact, combining the solutions generated from each of the sample future outcomes with the future outcome solutions generated by the AI Planner into an aggregated set of future outcome solutions, analyzing the aggregated set of future outcome solutions, selecting a best action based at least partially on the analysis of the aggregated set of future outcome solutions, and outputting the selected best action to computer memory.

Abstract: A data structure is described that comprises a balanced binary tree and a binary heap, which may be utilized for combinatorial searching algorithms. For instance, solutions for performing a task, such as a print job or the like, are associated with nodes that are utilized to generate the data structure. Each node is associated with a quality indicator that describes a most optimal solution that may be reached through the node when traversing the binary tree. The binary heap is generated from a subset of the nodes in the tree, wherein each node in the subset has a quality indicator value that is within a predefined range of a best known solution quality. The binary heap is sorted according to a search effort indicator value for each node, where nodes that are more easily reached in the tree are placed higher in the heap to facilitate rapid identification.

Abstract: Features described herein relate to optimizing a job plan procedure for selecting a plan for executing a manufacturing job. A planner can receive a model of a system that is to perform a job, and may select a precomputed plan that is sufficient to perform the job. The precomputed plan is identified during a first portion of a planning period, and the remainder of the planning period may be utilized to search for and identify a better plan for executing the job. If a better plan cannot be identified by the end of the planning period, then the selected precomputed plan can be executed.

Abstract: Respective coordinators are spawned or activated to coordinate activities with regard to respective tasks. Where the respective tasks require cooperative efforts of a plurality of controllers, the respective coordinators ensure cooperative efforts by generating and communicating cooperative commands to the plurality of controllers. The coordinators may act as clearinghouses for system data, selectively requesting and relaying system information to appropriate controllers. For example, a document processing system activates respective coordinators for respective sheets of print media. The respective coordinators orchestrate the transportation of the sheets by sequentially orchestrating the activities of sequentially selected pluralities of transportation actuator controllers. Selected sheet position information from sensors and/or from models maintained by the actuator controllers may be relayed by the coordinators to selected actuator controllers as appropriate to the sheet transportation tasks.

Abstract: A method for planning machine control for a system includes determining one or more capabilities and one or more capability constraints for each component used to execute a plan for processing a job by the system. The plan is incrementally constructed based on the components, the one or more capabilities, and the one or more constraints. One or more sets of executable instructions are queried with incremental portions of the plan, wherein each set of executable instructions is associated with a different one of the components and represents the actions that are performed by its corresponding component, each incremental portion includes actions that are to be performed by its corresponding component, and each set of executable instructions executes the incremental portion it received.

Abstract: Features described herein relate to concurrently processing multiple batches of job requests for one or more machines and/or components thereof, using a plurality of job planning queues. Each batch of job requests is allocated to a planning queue, and each planning queue comprises an unplanned subqueue that stores unplanned jobs, an unsent subqueue that stores planned jobs waiting to be executed, and a sent subqueue that stores planned jobs that have been output to the machine(s) for execution. A job planner and related components determine which unsent subqueue has the fewest planned jobs at a given point in time, and selects an unplanned job from the unplanned subqueue in the same planning queue as the identified unsent subqueue. The planner then generates a plan for the selected job and inserts the planned job into the unsent subqueue for eventual output to the machine(s) for execution.

Abstract: Features described herein relate to concurrently processing multiple batches of job requests for one or more machines and/or components thereof, using a plurality of job planning queues. Each batch of job requests is allocated to a planning queue, and each planning queue comprises an unplanned subqueue that stores unplanned jobs, an unsent subqueue that stores planned jobs waiting to be executed, and a sent subqueue that stores planned jobs that have been output to the machine(s) for execution. A job planner and related components determine which unsent subqueue has the fewest planned jobs at a given point in time, and selects an unplanned job from the unplanned subqueue in the same planning queue as the identified unsent subqueue. The planner then generates a plan for the selected job and inserts the planned job into the unsent subqueue for eventual output to the machine(s) for execution.

Abstract: A combinatorial search method and system is implemented in a computer control system for utilizing state-space planning of operations for multi-step production processes. The planner considers various possible combinations of actions, searching for one that correctly transforms the initial state of the object (or commodity) into the specified desired final state, where each combination of actions the planner considers is called a search node. Each node contains a plan representing a series of actions of a plurality of machines on a single object and also containing the predicted state of the object with those actions applied either forward or backward. The state of the object consists of the set of attributes of the object.

Abstract: A data structure is described that comprises a balanced binary tree and a binary heap, which may be utilized for combinatorial searching algorithms. For instance, solutions for performing a task, such as a print job or the like, are associated with nodes that are utilized to generate the data structure. Each node is associated with a quality indicator that describes a most optimal solution that may be reached through the node when traversing the binary tree. The binary heap is generated from a subset of the nodes in the tree, wherein each node in the subset has a quality indicator value that is within a predefined range of a best known solution quality. The binary heap is sorted according to a search effort indicator value for each node, where nodes that are more easily reached in the tree are placed higher in the heap to facilitate rapid identification.

Abstract: Features described herein relate to concurrently processing multiple batches of job requests for one or more machines and/or components thereof, using a plurality of job planning queues. Each batch of job requests is allocated to a planning queue, and each planning queue comprises an unplanned subqueue that stores unplanned jobs, an unsent subqueue that stores planned jobs waiting to be executed, and a sent subqueue that stores planned jobs that have been output to the machine(s) for execution. A job planner and related components determine which unsent subqueue has the fewest planned jobs at a given point in time, and selects an unplanned job from the unplanned subqueue in the same planning queue as the identified unsent subqueue. The planner then generates a plan for the selected job and inserts the planned job into the unsent subqueue for eventual output to the machine(s) for execution.

Abstract: Features described herein relate to optimizing a job plan procedure for selecting a plan for executing a manufacturing job. A planner can receive a model of a system that is to perform a job, and may select a precomputed plan that is sufficient to perform the job. The precomputed plan is identified during a first portion of a planning period, and the remainder of the planning period may be utilized to search for and identify a better plan for executing the job. If a better plan cannot be identified by the end of the planning period, then the selected precomputed plan can be executed.

Abstract: A method for planning machine control for a system includes determining one or more capabilities and one or more capability constraints for each component used to execute a plan for processing a job by the system. The plan is incrementally constructed based on the components, the one or more capabilities, and the one or more constraints. One or more sets of executable instructions are queried with incremental portions of the plan, wherein each set of executable instructions is associated with a different one of the components and represents the actions that are performed by its corresponding component, each incremental portion includes actions that are to be performed by its corresponding component, and each set of executable instructions executes the incremental portion it received.