As mobile applications grow in complexity, there is an increasing need to perform computationally intensive tasks. However, User Devices (UDs), such as tablets and smartphones, have limited capacity to carry out the required computations. Task offloading in Mobile Edge Computing (MEC) is a strategy that meets this demand by distributing tasks between UDs and servers. Deep Reinforcement Learning (DRL) is a promising solution for this strategy because it can adapt to dynamic changes and minimize online computational complexity. However, the combination of continuous-valued soft constraints and discrete-valued hard constraints on UDs and MEC servers poses significant challenges for designing efficient DRL algorithms. Existing DRL-based task-offloading algorithms focus on the constraints of the UDs, assuming the availability of enough resources on the server. Moreover, existing Multiagent DRL (MADRL)-based task-offloading algorithms are homogeneous agents and consider homogeneous constraints as a penalty in their reward function. We propose a novel Client–Master MADRL (CMMADRL) algorithm for task offloading in MEC that uses client agents at the UDs to decide on their resource requirements and a master agent at the server to make a combinatorial action selection based on the decision of the UDs. CMMADRL is shown to achieve up to 59% improvement in performance over existing benchmark and heuristic algorithms.

In modern urban transportation networks, multiple self-interested travel providers (public transit, micromobility providers, ride-sharing platforms and toll roads) compete for heterogenous transportation users that wish to balance time and cost. Traditional congestion models assume fixed, exogenous costs, while dynamic-pricing frameworks typically focus on a single operator, overlooking the rich strategic interplay among decentralised transportation providers. This paper introduces the Multi-Market Routing Problem (MMRP), a game-theoretic model in which each provider utilises adaptive pricing to maximise profit and heterogeneous transportation users aim to minimise their travel time and cost. We present the MMRP as an extension of traditional congestion games, and extend it to consider online instances for adaptive pricing under dynamic and stochastic congestion. We demonstrate the computational complexity for game-theoretic and exact solutions to the MMRP, reflecting the computational complexity of coordinating routing in dynamic and uncertain settings. To address this, we propose the use of independent Proximal Policy Optimisation as a decentralised and effective solution to the online MMRP, demonstrating reduced travel times and more equitable and fair outcomes for transportation users, and increased profitability for transportation providers. The MMRP framework and learning algorithms offer a principled foundation for competitive, multimodal routing in modern urban transportation networks.

This paper discusses the importance of ethical alignment in AI systems, particularly those designed with citizen end users in mind. It explores the intersection of responsible AI, socio-technical systems, and citizen-centric design, proposing that addressing the ethical aspect of decisions in citizen-centric AI systems enhances trust and acceptance of AI technologies. We focus on four key areas: (1) the formal specification of ethical principles, (2) processes to extract and elicit individual users’ ethical preferences, (3) aggregating these ethical preferences for a collective, and (4) mechanisms to ensure that the behaviour of AI systems aligns with the collective ethical preferences. We put forward a research roadmap by identifying challenges in these areas and highlighting solution concepts with the potential to address them.

Abstract In this paper, we address the problem of finding charging stops while travelling in electric vehicles (EVs) using artificial intelligence (AI). Choosing a charging station is challenging, because drivers have very heterogeneous preferences in terms of how they trade off the features of various alternatives (for example, regarding the time spent driving, charging costs, waiting times at charging stations, and the facilities provided at the charging stations). The key problem here is eliciting the diverse preferences of drivers, assuming that these preferences are typically not fully known a priori, and then planning stops based on each driver’s preferences. Our approach to solving this problem is to develop an intelligent personal agent that learns preferences gradually over multiple interactions. This study proposes a new technique that utilises a small-scale discrete choice experiment as a method of interacting with the driver in order to minimise the cognitive burden on the driver. Using this method, drivers are presented with a variety of routes with possible combinations of charging stops depending on the agent’s latest belief about their preferences. In subsequent iterations, the personal agent will continue to learn and refine its belief about the driver’s preferences, suggesting more personalised routes that are closer to the driver’s preferences. Based on real preference data from EV drivers, we evaluate our novel algorithm and show that, after only a few queries, our method quickly converges to the optimal routes for EV drivers [This paper is an extended version of an ECAI workshop short paper (Shafipour Yourdshahi et al., in: ECAI 2023 workshops, Kraków, Poland, 2023)].

Distinguishing epistemic from aleatoric uncertainty is a central idea to out-of-distribution (OOD) detection. By interpreting adversarial and OOD inputs from this perspective, we can collect them into a single unclassifiable group. Rejecting such inputs mid-inference will reduce resource usage. To achieve this, we apply k-nearest neighbour (KNN) classifiers to the embedding space of branched neural networks. This introduces a novel means of additional power savings, through an early-exit reject. Our technique works out-of-the-box on any branched neural network and can be competitive on OOD benchmarks, achieving an area under receiver operator characteristic (AUROC) of over 0.9 in most datasets, and scores of 0.95+ when identifying perturbed inputs. A mixed input test set is introduced, we show how OOD inputs can be identified up to 50% of the time, and adversarial inputs up to 85% of the time. In a balanced test environment, this equates to power savings of up to 18% in the OOD scenario and 40% in the adversarial scenario. This allows a more stringent in-distribution (ID) classification policy, leading to accuracy improvements of 15% and 20% on the OOD and adversarial tests, respectively, when compared to conventional exit policies operating under the same conditions.

In this paper, we introduce an audio-based framework for occupancy estimation, including a new public dataset, and evaluate occupancy in a ‘cocktail party’ scenario where the party is simulated by mixing audio to produce speech with overlapping talkers (1-10 people). To estimate the number of speakers in an audio clip, we explored five different types of speech signal features and trained several versions of our model using convolutional neural networks (CNNs). Further, we adapted the framework to be privacy-preserving by making random perturbations of audio frames in order to conceal speech content and speaker identity. We show that some of our privacy-preserving features perform better at occupancy estimation than original waveforms. We analyse privacy further using two adversarial tasks: speaker recognition and speech recognition. Our privacy-preserving models can estimate the number of speakers in the simulated cocktail party clips within 1-2 persons based on a mean-square error (MSE) of 0.9-1.6 and we achieve up to 34.9% classification accuracy while preserving speech content privacy. However, it is still possible for an attacker to identify individual speakers, which motivates further work in this area.

Abstract The success of crowdsourcing projects relies critically on motivating a crowd to contribute. One particularly effective method for incentivising participants to perform tasks is to run contests where participants compete against each other for rewards. However, there are numerous ways to implement such contests in specific projects, that vary in how performance is evaluated, how participants are rewarded, and the sizes of the prizes. Also, the best way to implement contests in a particular project is still an open challenge, as the effectiveness of each contest implementation (henceforth, incentive) is unknown in advance. Hence, in a crowdsourcing project, a practical approach to maximise the overall utility of the requester (which can be measured by the total number of completed tasks or the quality of the task submissions) is to choose a set of incentives suggested by previous studies from the literature or from the requester’s experience. Then, an effective mechanism can be applied to automatically select appropriate incentives from this set over different time intervals so as to maximise the cumulative utility within a given financial budget and a time limit. To this end, we present a novel approach to this incentive selection problem. Specifically, we formalise it as an online decision making problem, where each action corresponds to offering a specific incentive. After that, we detail and evaluate a novel algorithm, , to solve the incentive selection problem efficiently and adaptively. In theory, in the case that all the estimates in (except the estimates of the effectiveness of each incentive) are correct, we show that the algorithm achieves the regret bound of $\mathcal {O}(\sqrt {B/c})$ O ( B / c ) , where B denotes the financial budget and c is the average cost of the incentives. In experiments, the performance of is about 93% (up to 98%) of the optimal solution and about 9% (up to 40%) better than state-of-the-art algorithms in a broad range of settings, which vary in budget sizes, time limits, numbers of incentives, values of the standard deviation of the incentives’ utilities, and group sizes of the contests (i.e., the numbers of participants in a contest).

Edge computing has become a very popular service that enables mobile devices to run complex tasks with the help of network-based computing resources. However, edge clouds are often resource-constrained, which makes resource allocation a challenging issue. We focus on a distributed resource allocation method in which servers operate independently and do not communicate with each other, but interact with clients (tasks) to make allocation decisions. This provides robustness and does not require service providers to share information about their configurations or workloads. We utilize a two-round bidding approach of assigning tasks to edge cloud servers. We consider a preemption-enabled system in which servers may stop a previous task in order to run a more useful one. We evaluate the performance of our system using realistic simulations and real-world trace data from a high-performance computing cluster. Results show that our approach is reasonably close to optimal assignment, while saving 50–70 % of the original computation time.

Autonomous intersection management has the potential to reduce road traffic congestion and energy consumption. To realize this potential, efficient algorithms are needed. However, most existing studies locally optimize one intersection at a time, and this can cause negative externalities on the traffic network as a whole. Here, we focus on coordinating multiple intersections, and formulate the problem as a distributed constraint optimisation problem (DCOP). We consider three utility design approaches that trade off efficiency and fairness. Our polynomial-time algorithm for coordinating multiple intersections reduces the traffic delay by about 41 percentage points compared to independent single intersection management approaches.

Ridesharing is a promising approach for reducing congestion and pollution, and many variants have been studied in the literature over the past decades. In this paper, we consider a novel setting where individuals walk to a common pick-up point and ride together to a single drop-off point from where they walk to their final destination. This setting requires finding the optimal composition of riders and pick-up and drop-off meeting points, as well as an equitable distribution of the costs whereby riders are incentivised to participate. Based on game-theoretic principles, we propose a methodology to determine the optimal pick-up and drop-off points, together with a cost allocation method that is equitable in the sense that it ensures proportionality for sharing the costs, i.e., those who walk more should pay less. We present a formal evaluation of our cost allocation method and empirical evaluation against the Shapley value using real-world and simulated data. Our results show that our approach is computationally more tractable than the Shapley value, as it is linear in time while guaranteeing individual rationality under certain conditions.

The Internet of Vehicles aims to fundamentally improve transportation by connecting vehicles, drivers, passengers, and service providers together. Several new services such as parking space identification, platooning and intersection control—to name just a few—are expected to improve traffic congestion, reduce pollution, and improve the efficiency, safety and logistics of transportation. Proposed end-user services, however, make extensive use of private information with little consideration for the impact on users and third parties (those individuals whose information is indirectly involved). This article provides the first comprehensive overview of privacy and trust issues in the Internet of Vehicles at the service level. Various concerns over privacy are formalised into four basic categories: personal information privacy, multi-party privacy, trust, and consent to share information. To help analyse services and to facilitate future research, the main relevant end-user services are taxonomised according to voluntary and involuntary information they require and produce. Finally, this work identifies several open research problems and highlights general approaches to address them. These especially relate to measuring the trade-off between privacy and service functionality, automated consent negotiation, trust towards the IoV and its individual services, and identifying and resolving multi-party privacy conflicts.

Ridesharing has great potential to improve transportation efficiency while reducing congestion and pollution. To realize this potential, mechanisms are needed that allocate vehicles optimally and provide the right incentives to riders. However, many existing approaches consider restricted settings (e.g., only one rider per vehicle or a common origin for all riders). Moreover, naive applications of standard approaches, such as the Vickrey-Clarke-Groves or greedy mechanisms, cannot achieve a polynomial-time, truthful, individually rational and budget balanced mechanism. To address this, we formulate a general ridesharing problem and apply mechanism design to develop a novel mechanism which satisfies all four properties and whose social cost is within 8.6% of the optimal on average.

Edge computing has become a very popular service that enables mobile devices to run complex tasks with the help of network-based computing resources. However, edge clouds are often resource-constrained, which makes resource allocation a challenging issue. We focus on a distributed resource allocation method in which servers operate independently and do not communicate with each other, but interact with clients (tasks) to make allocation decisions. This provides robustness and does not require service providers to share information about their configurations or workloads. We propose a two-round bidding approach of assigning tasks to edge cloud servers, while taking into account various processing requirements and server constraints. We consider cases in which all jobs have equal utility, cases where jobs have different utilities but users do not disclose these utilities to servers, and cases where users disclose the utility of their jobs to servers. We evaluate the performance using extensive realistic simulations. Results show that our approach is very close to an optimal assignment, with discrepancy not exceeding 5%.

We identify optimal strategies for maximising influence within a social network in competitive settings under budget constraints. While existing work has focussed on simple threshold models, we consider more realistic settings, where (i) states are dynamic, i.e., nodes oscillate between influenced and uninfluenced states, and (ii) continuous amounts of resources (e.g., incentives or effort) can be expended on the nodes.

Computer vision and pattern recognition are increasingly being employed by smartphone and tablet applications targeted at lay-users. An open design challenge is to make such systems intelligible without requiring users to become technical experts. This paper reports a lab study examining the role of visual feedback. Our findings indicate that the stage of processing from which feedback is derived plays an important role in users' ability to develop coherent and correct understandings of a system's operation. Participants in our study showed a tendency to misunderstand the meaning being conveyed by the feedback, relating it to processing outcomes and higher level concepts, when in reality the feedback represented low level features. Drawing on the experimental results and the qualitative data collected, we discuss the challenges of designing interactions around pattern matching algorithms.

Mobile edge computing is an emerging technology to offer resource-intensive yet delay-sensitive applications from the edge of mobile networks, where a major challenge is to allocate limited edge resources to competing demands. While prior works often make a simplifying assumption that resources assigned to different users are non-sharable, this assumption does not hold for storage resources, where users interested in services (e.g., data analytics) based on the same set of data/code can share storage resource. Meanwhile, serving each user request also consumes non-sharable resources (e.g., CPU cycles, bandwidth). We study the optimal provisioning of edge services with non-trivial demands of both sharable (storage) and non-sharable (communication, computation) resources via joint service placement and request scheduling. In the homogeneous case, we show that while the problem is polynomial-time solvable without storage constraints, it is NP-hard even if each edge cloud has unlimited communication or computation resources. We further show that the hardness is caused by the service placement subproblem, while the request scheduling subproblem is polynomial-time solvable via maximum-flow algorithms. In the general case, both subproblems are NP-hard. We develop a constant-factor approximation algorithm for the homogeneous case and efficient heuristics for the general case. Our trace-driven simulations show that the proposed algorithms, especially the approximation algorithm, can achieve near-optimal performance, serving 2-3 times more requests than a baseline solution that optimizes service placement and request scheduling separately.

Environmental monitoring allows authorities to understand the impact of potentially harmful phenomena, such as air pollution, excessive noise, and radiation. Recently, there has been considerable interest in participatory sensing as a paradigm for such large-scale data collection because it is cost-effective and able to capture more fine-grained data than traditional approaches that use stationary sensors scattered in cities. In this approach, ordinary citizens (non-expert contributors) collect environmental data using low-cost mobile devices. However, these participants are generally self-interested actors that have their own goals and make local decisions about when and where to take measurements. This can lead to highly inefficient outcomes, where observations are either taken redundantly or do not provide sufficient information about key areas of interest. To address these challenges, it is necessary to guide and to coordinate participants, so they take measurements when it is most informative. To this end, we develop a computationally-efficient coordination algorithm (adaptive Best-Match) that suggests to users when and where to take measurements. Our algorithm exploits probabilistic knowledge of human mobility patterns, but explicitly considers the uncertainty of these patterns and the potential unwillingness of people to take measurements when requested to do so. In particular, our algorithm uses a local search technique, clustering and random simulations to map participants to measurements that need to be taken in space and time. We empirically evaluate our algorithm on a real-world human mobility and air quality dataset and show that it outperforms the current state of the art by up to 24% in terms of utility gained.

In this article, we address the interrelated challenges of predicting user comfort and using this to reduce energy consumption in smart heating, ventilation, and air conditioning (HVAC) systems. At present, such systems use simple models of user comfort when deciding on a set-point temperature. Being built using broad population statistics, these models generally fail to represent individual users’ preferences, resulting in poor estimates of the users’ preferred temperatures. To address this issue, we propose the Bayesian Comfort Model (BCM). This personalised thermal comfort model uses a Bayesian network to learn from a user’s feedback, allowing it to adapt to the users’ individual preferences over time. We further propose an alternative to the ASHRAE 7-point scale used to assess user comfort. Using this model, we create an optimal HVAC control algorithm that minimizes energy consumption while preserving user comfort. Through an empirical evaluation based on the ASHRAE RP-884 dataset and data collected in a separate deployment by us, we show that our model is consistently 13.2% to 25.8% more accurate than current models and how using our alternative comfort scale can increase our model’s accuracy. Through simulations we show that using this model, our HVAC control algorithm can reduce energy consumption by 7.3% to 13.5% while decreasing user discomfort by 24.8% simultaneously.

Participatory sensing (PS) has gained significant attention as a crowdsourcing methodology that allows ordinary citizens (non-expert contributors) to collect data using low-cost mobile devices. In particular, it has been useful in the collection of environmental data. However, current PS applications suffer from two problems. First, they do not coordinate the measurements taken by their users, which is required to maximise system efficiency. Second, they are vulnerable to malicious behaviour. In this context, we propose a novel algorithm that simultaneously addresses both of these problems. Specifically, we use heteroskedastic Gaussian Processes to incorporate users' trustworthiness into a Bayesian spatio-temporal regression model. The model is trained with measurements taken by participants, thus it is able to estimate the value of the phenomenon at any spatio-temporal location of interest and also learn the level of trustworthiness of each user. Given this model, the coordination system is able to make informed decisions concerning when, where and who should take measurements over a period of time. We empirically evaluate our algorithm on a real-world human mobility and air quality dataset, where malicious behaviour is synthetically produced, and show that our algorithm outperforms the current state of the art by up to 60.4% in terms of RMSE while having a reasonable runtime.

We consider settings where electric vehicle drivers participate in a market mechanism to charge their vehicles. Existing work typically assumes that participants are fully rational and can report their charging preferences accurately. However, this may not be reasonable in settings with non-experts. To explore this, we design a novel game called Bid2Charge and compare a fully expressive interface that covers the entire space of preferences to two restricted interfaces that offer fewer possible reports. We show that restricting the users' preferences significantly reduces deliberation times while also leading to an increase in utility by up to 70%.

We consider settings where owners of electric vehicles (EVs) participate in a market mechanism to charge their vehicles. Existing work on such mechanisms has typically assumed that participants are fully rational and can report their preferences accurately via some interface to the mechanism or to a software agent participating on their behalf. However, this may not be reasonable in settings with non-expert human end-users.Thus, our overarching aim in this paper is to determine experimentally if a fully expressive market interface that enables accurate preference reports is suitable for the EV charging domain, or, alternatively, if a simpler, restricted interface that reduces the space of possible options is preferable. In doing this, we measure the performance of an interface both in terms of how it helps participants maximise their utility and how it affects deliberation time. Our secondary objective is to contrast two different types of restricted interfaces that vary in how they restrict the space of preferences that can be reported. To enable this analysis, we develop a novel game that replicates key features of an abstract EV charging scenario. In two experiments with over 300 users, we show that restricting the users' preferences significantly reduces the time they spend deliberating (by up to half in some cases). An extensive usability survey confirms that this restriction is furthermore associated with a lower perceived cognitive burden on the users. More surprisingly, at the same time, using restricted interfaces leads to an increase in the users' performance compared to the fully expressive interface (by up to 70%). We also show that some restricted interfaces have the desirable effect of reducing the energy consumption of their users by up to 20% while achieving the same utility as other interfaces. Finally, we find that a reinforcement learning agent displays similar performance trends to human users, enabling a novel methodology for evaluating market interfaces.

Participatory sensing is a promising new low-cost approach for collecting environmental data. However, current large-scale environmental participatory sensing campaigns typically do not coordinate the measurements of participants, which can lead to gaps or redundancy in the collected data. While some work has considered this problem, it has made several unrealistic assumptions. In particular, it assumes that complete and accurate knowledge about the participants future movements is available and it does not consider constraints on the number of measurements a user is willing to take. To address these shortcomings, we develop a computationally-efficient coordination algorithm (Best-match) to suggest to users where and when to take measurements. Our algorithm exploits human mobility patterns, but explicitly considers the inherent uncertainty of these patterns. We empirically evaluate our algorithm on a real-world human mobility and air quality dataset and show that it outperforms the state-of-the-art greedy and pull-based proximity algorithms in dynamic environments.

This paper introduces a novel intention-aware routing system (IARS) for electric vehicles. This system enables vehicles to compute a routing policy that minimizes their expected journey time while considering the policies, or intentions, of other vehicles. Considering such intentions is critical for electric vehicles, which may need to recharge en route and face potentially significant queueing times if other vehicles choose the same charging stations. To address this, the computed routing policy takes into consideration predicted queueing times at the stations, which are derived from the current intentions of other electric vehicles. The efficacy of IARS is demonstrated through simulations using realistic settings based on real data from The Netherlands, including charging station locations, road networks, historical travel times, and journey origin-destination pairs. In these settings, IARS is compared with a number of state-of-the-art benchmark routing algorithms and achieves significantly lower average journey times. In some cases, IARS leads to an over 80% improvement in waiting times at charging stations and a more than 50% reduction in overall journey times.

Finding and tracking targets and events in a live video feed is important for many commercial applications, from CCTV surveillance used by police and security firms, to the rapid mapping of events from aerial imagery. However, descriptions of targets are typically provided in natural language by the end users, and interpreting these in the context of a live video stream is a complex task. Due to current limitations in artificial intelligence, especially vision, this task cannot be automated and instead requires human supervision. Hence, in this paper, we consider the use of real-time crowdsourcing to identify and track targets given by a natural language description. In particular we present a novel method for augmenting live video with a real-time crowd.

Word-of-mouth, referral, or viral marketing is a highly sought-after way of advertising. In this paper, we investigate whether such marketing can be encouraged through incentive mechanisms, thus allowing an organisation to effectively crowdsource their marketing. Specifically, we undertake a field experiment that compares several mechanisms for incentivising social media shares in support of a charitable cause. Our experiment takes place on a website promoting a fundraising drive by a large cancer research charity. Site visitors who sign up to support the cause are asked to spread the word about it on Facebook, Twitter or other channels. They are randomly assigned to one of four treatments that differ in the way social sharing activities are incentivised. Under the control treatment, no extra incentive is provided. Under two of the other mechanisms, the sharers are offered a fixed number of points that help take the campaign further. We compare low and high levels of such incentives for direct referrals. In the final treatment, we adopt a multi-level incentive mechanism that rewards direct as well as indirect referrals (where referred contacts refer others). We find that providing a high level of incentives results in a statistically significant increase in sharing behaviour and resulting signups. Our data does not indicate a statistically significant increase for the low and multi-level incentive mechanisms.

We develop an online mechanism for the allocation of an expiring resource to a dynamic agent population. Each agent has a non-increasing marginal valuation function for the resource, and an upper limit on the number of units that can be allocated in any period. We propose two versions on a truthful allocation mechanism. Each modifies the decisions of a greedy online assignment algorithm by sometimes cancelling an allocation of resources. One version makes this modification immediately upon an allocation decision while a second waits until the point at which an agent departs the market. Adopting a prior-free framework, we show that the second approach has better worst-case allocative efficiency and is more scalable. On the other hand, the first approach (with immediate cancellation) may be easier in practice because it does not need to reclaim units previously allocated. We consider an application to recharging plug-in hybrid electric vehicles (PHEVs). Using data from a real-world trial of PHEVs in the UK, we demonstrate higher system performance than a fixed price system, performance comparable with a standard, but non-truthful scheduling heuristic, and the ability to support 50% more vehicles at the same fuel cost than a simple randomized policy.

In this paper, we develop a novel algorithm that allows service consumers to execute business processes (or workflows) of interdependent services in a dependable manner within tight time-constraints. In particular, we consider large interorganizational service-oriented systems, where services are offered by external organizations that demand financial remuneration and where their use has to be negotiated in advance using explicit service-level agreements (as is common in Grids and cloud computing). Here, different providers often offer the same type of service at varying levels of quality and price. Furthermore, some providers may be less trustworthy than others, possibly failing to meet their agreements. To control this unreliability and ensure end-to-end dependability while maximizing the profit obtained from completing a business process, our algorithm automatically selects the most suitable providers. Moreover, unlike existing work, it reasons about the dependability properties of a workflow, and it controls these by using service redundancy for critical tasks and by planning for contingencies. Finally, our algorithm reserves services for only parts of its workflow at any time, in order to retain flexibility when failures occur. We show empirically that our algorithm consistently outperforms existing approaches, achieving up to a 35-fold increase in profit and successfully completing most workflows, even when the majority of providers fail.

Web services promise to revolutionize the way computational resources and business processes are offered and invoked in open, distributed systems, such as the Internet. These services are described using machine-readable metadata, which enables consumer applications to automatically discover and provision suitable services for their workflows at run-time. However, current approaches have typically assumed service descriptions are accurate and deterministic, and so have neglected to account for the fact that services in these open systems are inherently unreliable and uncertain. Specifically, network failures, software bugs and competition for services may regularly lead to execution delays or even service failures. To address this problem, the process of provisioning services needs to be performed in a more flexible manner than has so far been considered, in order to proactively deal with failures and to recover workflows that have partially failed. To this end, we devise and present a heuristic strategy that varies the provisioning of services according to their predicted performance. Using simulation, we then benchmark our algorithm and show that it leads to a 700% improvement in average utility, while successfully completing up to eight times as many workflows as approaches that do not consider service failures.