Lineage is the process of keeping track of the relationship between the inputs of a data processing task and the parts of the output they contribute to produce. Depending on its precise definition, lineage can be seen as a form of database provenance, a means of tracking information flow in computer programs, or be used to express causality and provide counter-examples for the falsity of a logical statement. In this paper, we establish the formal foundations of a notion of lineage for arbitrary abstract functions manipulating objects that are “composite” –that is, can be made of multiple other objects. Three definitions of lineage over functions are formally defined, respectively called explanation, participation and extraction; we then establish explanation relationships for a set of elementary functions , and for compositions thereof. A fully functional implementation of these concepts is finally presented and experimentally evaluated.

Over the past years, Robotic Process Automation (RPA) has emerged as a significant tool to enhance productivity across various industries by automating repetitive tasks performed on computer user interfaces, thereby reducing error rates. In this paper, the RPA problem is addressed as an unbounded assignment problem. Specifically, the objective of the problem is to assign robots to transactions that must be completed at specific time periods, while minimizing the total number of robots required. The problem is solved using a bipartite graph representation and a random walk approximation, which defines an ordering of the transactions and periods in order to determine a valid assignment. The heuristic is evaluated on a real data set from a financial institution and compared to previous results obtained on generated data. The results obtained with the random walk approximation heuristics on the real data set is optimal in terms of number of robots.

Robotic Process Automation has emerged in recent years as an important field by allowing faster and more secure processes through a reduction in the risks or errors but also an increase in the productivity rates of many industries. In this specific paper, the RPA problem aims at assigning financial transactions to software robots to minimize the total costs, induced by the licenses and utilization time. The problem is represented as a bipartite graph and the effective resistance of the graph, which is analog to an electrical circuit, is used to order the edges of a heuristic method to assign the transactions to robots. Preliminary results, based on real data from a bank, are compared to the optimal solution obtained by a linear integer programming model. They show that the heuristic method allows to obtain results quicker and that they are near the optimal solution.

Abstract Explainability is the process of linking part of the inputs given to a calculation to its output, in such a way that the selected inputs somehow “cause” the result. We establish the formal foundations of a notion of explainability for arbitrary abstract functions manipulating nested data structures. We then establish explanation relationships for a set of elementary functions, and for compositions thereof. A fully functional implementation of these concepts is finally presented and experimentally evaluated.

Robotic Process Automation (RPA) is used in various fields of human activity in order to implement faster and more secure processes through a reduction in the risks or errors but also an increase in the productivity rates. The increase of its use and importance calls for evermore efficient solution methods for this problem. In this paper, the RPA is addressed in the context of a financial institution. The problem consists in assigning transactions to software robots, where each transaction type has a different clearance date and a different processing time. First, four heuristics are used to compute an upper bound on the number of required software robots. Then, this bound is given as a parameter to an integer linear program, which is used to assign the transactions to the different robots. The quality of the solutions is assessed by an extensive experimental study on a set of 39,000 instances. The results show that two heuristics outperform the others and allow for a faster resolution by the integer linear program which in turn finds the optimal solution for most of the instances within a timeout of 60 seconds.

Robotic Process Automation (RPA) is used in various fields of human activity in order to implement faster and more secure processes through a reduction in the risks or errors but also an increase in the productivity rates. The increase of its use and importance calls for evermore efficient solution methods for this problem. In this paper, the RPA is addressed in the context of a financial institution. The problem consists in assigning transactions to software robots, where each transaction type has a different clearance date and a different processing time. First, four heuristics are used to compute an upper bound on the number of required software robots. Then, this bound is given as a parameter to an integer linear program, which is used to assign the transactions to the different robots. The quality of the solutions is assessed by an extensive experimental study on a set of 39,000 instances. The results show that two heuristics outperform the others and allow for a faster resolution by the integer linear program which in turn finds the optimal solution for most of the instances within a timeout of 60 seconds.

Robotic process automation (RPA) helps companies reduce the time required to process tasks by using software or robots to mimic human actions on graphic interfaces. In this paper, the RPA problem is solved for a financial institution. A set of different types of financial transactions are to be processed with different processing times, volumes, market hours and clearance delays. In a previous work, a two-phase linear integer model was used to solve the problem on small instances. In this study, a network flow algorithm is used to compute a lower bound for the problem, thus reducing the computational time required to obtain a solution. The method is tested on a real case provided by a bank in North America and on synthetic test cases containing a greater number of transaction types. Results show that combining the computation of the lower bound with a linear integer model is faster and more practical.