Autonomous robots are agents that interact with the environment and perform tasks using their own abilities (i.e., skills) without continuous human intervention. However, in real-life scenarios, intelligent robots also need to discover the effects of their actions and understand how to save them for future use. This task appears time-consuming and very challenging, especially in a social environment populated by people who typically modify their behaviors based on the context and can dynamically impact the robot’s decision-making process. This paper aims to investigate the feasibility of autonomously creating an abstract representation of the domain knowledge from the data acquired during the robot’s exploration, inferring causal-effect relations between the executed actions, and learning context-aware symbols that describe the environment states at high level, ultimately producing a PDDL-based description of the domain. With this purpose, a new framework that relies on ROS, the standard de-facto in robotics, and ROSPlan has been developed to facilitate the transfer into several robotic platforms. Preliminary results suggest the possibility of describing the robot’s experience per option via context-based symbols that are consistently learned by the system from a few data samples.

Abstract The gate-based model is one of the leading quantum computing paradigms for representing quantum circuits. Within this paradigm, a quantum algorithm is expressed in terms of a set of quantum gates that are executed on the quantum hardware over time, subject to a number of constraints whose satisfaction must be guaranteed before running the circuit, to allow for feasible execution. The need to guarantee the previous feasibility condition gives rise to the Quantum Circuit Compilation Problem (QCCP). The QCCP has been demonstrated to be NP-Complete, and can be considered as a Planning and Scheduling problem. In this paper, we consider quantum compilation instances deriving from the general Quantum Approximation Optimization Algorithm (QAOA), applied to the MaxCut problem, devised to be executed on Noisy Intermediate Scale Quantum (NISQ) hardware architectures. More specifically, in addition to the basic QCCP version, we also tackle other variants of the same problem such as the QCCP-X (QCCP with crosstalk constraints), the QCCP-V (QCCP with variable qubit state initialization), as well as the QCCP-VX that includes both previous variants. All problem variants are solved using genetic algorithms. We perform an experimental study across a conventional set of instances taken from the literature, and show that the proposed genetic algorithm, termed $$GA_{VX}$$ G A VX , outperforms previous approaches in the literature.

Energy awareness is one of the most relevant research directions in scheduling problems. In this paper we consider the minimization of both the makespan and the energy consumption in the classical job shop scheduling problem. The energy model considered allows several possible states for the machines: off, stand-by, idle, setup and processing. To solve this multi-objective problem we propose an NSGA-II based evolutionary algorithm combined with local search and a heuristic procedure to improve the energy consumption of a given schedule. We also propose an advanced constraint programming (CP) approach as well as a Mixed-Integer Linear Programming (MILP) model, to the aim of comparing their performances against those obtained with the NSGA-II. The experimental study is performed against a benchmark set that extends by 41 instances of increasing size, the set tackled in the previous literature against the same problem. The experiments demonstrate the superiority of the NSGA-II algorithm over all other methods, despite the utilization of CP and MILP allows to draw interesting conclusions on the overall solution optimality, revealing that there is still room for further optimization.

This paper presents the achievements obtained from a study performed within the IMPACT (Intrinsically Motivated Planning Architecture for Curiosity-driven roboTs) Project funded by the European Space Agency (ESA).The main contribution of the work is the realization of an innovative robotic architecture in which the well-known three-layered architectural paradigm (decisional, executive, and functional) for controlling robotic systems is enhanced with autonomous learning capabilities.The architecture is the outcome of the application of an interdisciplinary approach integrating Artificial Intelligence (AI), Autonomous Robotics, and Machine Learning (ML) techniques.In particular, state-of-the-art AI planning systems and algorithms were integrated with Reinforcement Learning (RL) algorithms guided by intrinsic motivations (curiosity, exploration, novelty, and surprise).The aim of this integration was to: (i) develop a software system that allows a robotic platform to autonomously represent in symbolic form the skills autonomously learned through intrinsic motivations; (ii) show that the symbolic representation can be profitably used for automated planning purposes, thus improving the robot's exploration and knowledge acquisition capabilities.The proposed solution is validated in a test scenario inspired by a typical space exploration mission involving a rover.

In this work we investigate the performance of greedy randomised search (GRS) techniques to the problem of compiling quantum circuits to emerging quantum hardware. Quantum computing (QC) represents the next big step towards power consumption minimisation and CPU speed boost in the future of computing machines. Quantum computing uses quantum gates that manipulate multi-valued bits (qubits). A quantum circuit is composed of a number of qubits and a series of quantum gates that operate on those qubits, and whose execution realises a specific quantum algorithm. Current quantum computing technologies limit the qubit interaction distance allowing the execution of gates between adjacent qubits only. This has opened the way to the exploration of possible techniques aimed at guaranteeing nearest-neighbor (NN) compliance in any quantum circuit through the addition of a number of so-called swap gates between adjacent qubits. In addition, technological limitations (decoherence effect) impose that the overall duration (makespan) of the quantum circuit realization be minimized. One core contribution of the paper is the definition of two lexicographic ranking functions for quantum gate selection, using two keys: one key acts as a global closure metric to minimise the solution makespan; the second one is a local metric, which favours the mutual approach of the closest qstates pairs. We present a GRS procedure that synthesises NN-compliant quantum circuits realizations, starting from a set of benchmark instances of different size belonging to the Quantum Approximate Optimization Algorithm (QAOA) class tailored for the MaxCut problem. We propose a comparison between the presented meta-heuristics and the approaches used in the recent literature against the same benchmarks, both from the CPU efficiency and from the solution quality standpoint. In particular, we compare our approach against a reference benchmark initially proposed and subsequently expanded in [1] by considering: (i) variable qubit state initialisation and (ii) crosstalk constraints that further restrict parallel gate execution.

Quantum Computing represents the next big step towards speed boost in computation, which promises major breakthroughs in several disciplines including Artificial Intelligence. This paper investigates the performance of a genetic algorithm to optimize the realization (compilation) of nearest-neighbor compliant quantum circuits. Currrent technological limitations (e.g., decoherence effect) impose that the overall duration (makespan) of the quantum circuit realization be minimized, and therefore the makespanminimization problem of compiling quantum algorithms on present or future quantum machines is dragging increasing attention in the AI community. In our genetic algorithm, a solution is built utilizing a novel chromosome encoding where each gene controls the iterative selection of a quantum gate to be inserted in the solution, over a lexicographic double-key ranking returned by a heuristic function recently published in the literature.Our algorithm has been tested on a set of quantum circuit benchmark instances of increasing sizes available from the recent literature. We demonstrate that our genetic approach obtains very encouraging results that outperform the solutions obtained in previous research against the same benchmark, succeeding in significantly improving the makespan values for a great number of instances.

One of the most recent and interesting trends in intelligent scheduling is trying to reduce the energy consumption in order to obtain lower production costs and smaller carbon footprint. In this work we consider the energy-aware job shop scheduling problem, where we have to minimize at the same tim e an efficiency-based objective, as is the total weighted tardiness, and also the overall energy consumption. We experimentally show that we can reduce the energy consumption of a given schedule by delaying some operations, and to this end we design a heuristic procedure to improve a given schedule. As the problem is computationally complex, we design three approaches to solve it: a Pareto-based multiobjective evolutionary algorithm, which is hybridized with a multiobjective local search method and a linear programming step, a decomposition-based multiobjective evolutionary algorithm hybridized with a single-objective local search method, and finally a constraint programming approach. We perform an extensive experimental study to analyze our algorithms and to compare them with the state of the art.

In this paper, we tackle the Energy-Flexible Flow Shop Scheduling (EnFFS) problem, a multi-objective optimization problem focused on the minimization of both the overall completion time C and the global energy consumption E of the solutions. The tackled problem is an extension of the Flexible Flow-Shop Scheduling problem where each activity in a job has a set of possible execution modes with different trade-off between energy consumed and processing time. Moreover, global energy consumption E may also depend on the possibility to switch-off the machines during the idle periods. The goal of this work is to widen the knowledge about performance capabilities, in particular the ability of efficiently finding high quality approximations of the solution Pareto front. To this aim, we explore the development of innovative meta-heuristic algorithms for solving the proposed multi-objective scheduling problem. In particular, we consider stochastic local search (SLS) algorithms, introducing a Multi-Objective Large Neighbourhood Search (MO-LNS) framework in line with the large neighbourhood search approaches proposed in literature, and present some preliminary results obtained against a EnFFS benchmark recently proposed in the literature, showing some initial but appreciable improvements. Moreover, four new instances are presented and experimented upon, which will be hopefully used by the community as a novel benchmark on which to test new multi-objective optimisation research contributions.

This article elaborates around a recent effort to build a planning system that helps human mission planning in a space mission. Specifically, the article describes the steps that brought us to develop MrSPOCK, the MARS EXPRESS Science Plan Opportunities Coordination Kit, a tool for supporting long‐term planning in the MARS EXPRESS mission of the European Space Agency. In showing our effort for creating MrSPOCK, we will underscore the key ingredients for developing complete applications and how they are connected to a stable line of research on planning and scheduling with timelines.

The main result of the paper is the reduction of the RCPSP/max problem to a Disjunctive Temporal Problem that allows customization of specific properties within a backtracking search procedure for makespan optimization. In addition, a branching strategy is proposed able to deduce new constraints which explicitly represent infeasible or useless search paths. A new variable ordering heuristic (called clustering) is also used which provides a further boosting to the algorithm's effectiveness.

This paper describes RAXEM, an AI-based system developed to support human mission planners in the daily task to plan uplink commands for an interplanetary spacecraft. The intelligent environment of RAXEM has been designed to support the users in analyzing the problem and taking planning decisions as a result of an interactive process. The system combines different ingredients like integrating flexible automated algorithms, promoting user active participation during problem solving, and guaranteeing continuity of work practice. The paper touches upon all these aspects and comments on how a key factor for success has been the integration of intelligent technology to continuously support mission plan management.

In the realm of scheduling problems, different sources of uncertainty can invalidate the planned solutions: unpredictability of activity behaviors, machine breakdowns, new activities to be served, and so on. In this paper, we are concerned with the generation of high-quality downlink schedules in a spacecraft domain in the presence of a high degree of uncertainty. In particular, we refer to a combinatorial optimization problem called Mars-Express memory dumping problem (MEX-MDP), which arose in the European Space Agency (ESA) program Mars-Express. A MEX-MDP consists of a generation of dumping commands to maximize the downloads of data sets from the satellite to the ground. The domain is characterized with several kinds of constraints - such as bounded onboard memory capacities, limited communication windows over the downlink channels, deadlines, and ready times imposed by the payload requirements - and different sources of uncertainty - such as the amount of data generated at each scientific observation or the channel data rate. In this paper, we tackle this problem by using a reduction of the MEX-MDP to a max-flow problem: the former problem has a solution when the maximum flow in the latter equates the total amount of data to dump. Given this reduction, we introduce a novel definition of solution robustness based on the utilization of the onboard memory, as well as an interative procedure to improve solution quality. The key idea behind this approach is that the lower the in memory utilization, the higher the ability to cope with an unexpectedly large amount of data.

Recently, the areas of planning and scheduling in artificial intelligence (AI) have witnessed a big push toward their integration in order to solve complex problems. These problems require both reasoning on which actions are to be performed as well as their precedence constraints (planning) and the reasoning with respect to temporal constraints (e.g., duration, precedence, and deadline); those actions should satisfy the resources they use (scheduling). This paper describes IPSS (integrated planning and scheduling system), a domain independent solver that integrates an AI planner that synthesizes courses of actions with constraint-based techniques that reason based upon time and resources. IPSS is able to manage not only simple precedence constraints, but also more complex temporal requirements (as the Allen primitives) and multicapacity resource usage/consumption. The solver is evaluated against a set of problems characterized by the use of multiple agents (or multiple resources) that have to perform tasks with some temporal restrictions in the order of the tasks or some constraints in the availability of the resources. Experiments show how the integrated reasoning approach improves plan parallelism and gains better makespans than some state-of-the-art planners where multiple agents are represented as additional fluents in the problem operators. It also shows that IPSS is suitable for solving real domains (i.e., workflow problems) because it is able to impose temporal windows on the goals or set a maximum makespan, features that most of the planners do not yet incorporate

A module for quantitative temporal management is presented which can be easily connected with different problem solving architectures. Particular attention has been devoted to planning and scheduling problems in realistic domains. The module allows for an incremental constraint posting approach for building solutions. After introducing the consistency-checking problem in quantitative temporal networks, a correct and complete algorithm for constraint propagation is described, and a sufficient condition for inconsistency is also presented as useful to improve the algorithm. Moreover a repropagate operator is given which is used to re-establish a consistent network when a given constraint set is relaxed.