Search

Distribution factors are indispensable tools in the design and analysis of power transmission grids. Recently, they received a renewed interest in the field of topology optimization, leading to the definition of bus merge and bus split distribution factors. In this article, we introduce a unified derivation of the most relevant distribution factors based on matrix algebraic manipulations. This approach facilitates the generalization to more complex grid modification, in particular simultaneous switching events or bus splits., Comment: 11 pages, 1 figure

Computational modelling of materials using machine learning, ML, and historical data has become integral to materials research. The efficiency of computational modelling is strongly affected by the choice of the numerical representation for describing the composition, structure and chemical elements. Structure controls the properties, but often only the composition of a candidate material is available. Existing elemental descriptors lack direct access to structural insights such as the coordination geometry of an element. In this study, we introduce Local Environment-induced Atomic Features, LEAFs, which incorporate information about the statistically preferred local coordination geometry for atoms in crystal structure into descriptors for chemical elements, enabling the modelling of materials solely as compositions without requiring knowledge of their crystal structure. In the crystal structure, each atomic site can be described by similarity to common local structural motifs; by aggregating these features of similarity from the experimentally verified crystal structures of inorganic materials, LEAFs formulate a set of descriptors for chemical elements and compositions. The direct connection of LEAFs to the local coordination geometry enables the analysis of ML model property predictions, linking compositions to the underlying structure-property relationships. We demonstrate the versatility of LEAFs in structure-informed property predictions for compositions, mapping of chemical space in structural terms, and prioritising elemental substitutions. Based on the latter for predicting crystal structures of binary ionic compounds, LEAFs achieve the state-of-the-art accuracy of 86 per cent. These results suggest that the structurally informed description of chemical elements and compositions developed in this work can effectively guide synthetic efforts in discovering new materials., Comment: 10 pages, 4 figures, supplementary information

Random projection, a dimensionality reduction technique, has been found useful in recent years for reducing the size of optimization problems. In this paper, we explore the use of sparse sub-gaussian random projections to approximate semidefinite programming (SDP) problems by reducing the size of matrix variables, thereby solving the original problem with much less computational effort. We provide some theoretical bounds on the quality of the projection in terms of feasibility and optimality that explicitly depend on the sparsity parameter of the projector. We investigate the performance of the approach for semidefinite relaxations appearing in polynomial optimization, with a focus on combinatorial optimization problems. In particular, we apply our method to the semidefinite relaxations of MAXCUT and MAX-2-SAT. We show that for large unweighted graphs, we can obtain a good bound by solving a projection of the semidefinite relaxation of MAXCUT. We also explore how to apply our method to find the stability number of four classes of imperfect graphs by solving a projection of the second level of the Lasserre Hierarchy. Overall, our computational experiments show that semidefinite programming problems appearing as relaxations of combinatorial optimization problems can be approximately solved using random projections as long as the number of constraints is not too large.

We propose DFAMiner, a passive learning tool for learning minimal separating deterministic finite automata (DFA) from a set of labelled samples. Separating automata are an interesting class of automata that occurs generally in regular model checking and has raised interest in foundational questions of parity game solving. We first propose a simple and linear-time algorithm that incrementally constructs a three-valued DFA (3DFA) from a set of labelled samples given in the usual lexicographical order. This 3DFA has accepting and rejecting states as well as don't-care states, so that it can exactly recognise the labelled examples. We then apply our tool to mining a minimal separating DFA for the labelled samples by minimising the constructed automata via a reduction to solving SAT problems. Empirical evaluation shows that our tool outperforms current state-of-the-art tools significantly on standard benchmarks for learning minimal separating DFAs from samples. Progress in the efficient construction of separating DFAs can also lead to finding the lower bound of parity game solving, where we show that DFAMiner can create optimal separating automata for simple languages with up to 7 colours. Future improvements might offer inroads to better data structures., Comment: 24 pages including appendices and references; version for LearnAut workshop

Climate risk assessments must account for a wide range of possible futures, so scientists often use simulations made by numerous global climate models to explore potential changes in regional climates and their impacts. Some of the latest-generation models have high effective climate sensitivities or EffCS. It has been argued these so-called hot models are unrealistic and should therefore be excluded from analyses of climate change impacts. Whether this would improve regional impact assessments, or make them worse, is unclear. Here we show there is no universal relationship between EffCS and projected changes in a number of important climatic drivers of regional impacts. Analysing heavy rainfall events, meteorological drought, and fire weather in different regions, we find little or no significant correlation with EffCS for most regions and climatic drivers. Even when a correlation is found, internal variability and processes unrelated to EffCS have similar effects on projected changes in the climatic drivers as EffCS. Model selection based solely on EffCS appears to be unjustified and may neglect realistic impacts, leading to an underestimation of climate risks., Comment: Preprint, 30 pages, 4 figures and 2 tables

While discounted payoff games and classic games that reduce to them, like parity and mean-payoff games, are symmetric, their solutions are not. We have taken a fresh view on the properties that optimal solutions need to have, and devised a novel way to converge to them, which is entirely symmetric. We achieve this by building a constraint system that uses every edge to define an inequation, and update the objective function by taking a single outgoing edge for each vertex into account. These edges loosely represent strategies of both players, where the objective function intuitively asks to make the inequation to these edges sharp. In fact, where they are not sharp, there is an `error' represented by the difference between the two sides of the inequation, which is 0 where the inequation is sharp. Hence, the objective is to minimise the sum of these errors. For co-optimal strategies, and only for them, it can be achieved that all selected inequations are sharp or, equivalently, that the sum of these errors is zero. While no co-optimal strategies have been found, we step-wise improve the error by improving the solution for a given objective function or by improving the objective function for a given solution. This also challenges the gospel that methods for solving payoff games are either based on strategy improvement or on value iteration., Comment: arXiv admin note: substantial text overlap with arXiv:2310.01008

The verification of liveness conditions is an important aspect of state-based rigorous methods. This article investigates this problem in a fragment $\square$LTL of the logic LTL(EB), the integration of the UNTIL-fragment of Pnueli's linear time temporal logic (LTL) and the logic of Event-B, in which the most commonly used liveness conditions can be expressed. For this fragment a sound set of derivation rules is developed, which is also complete under mild restrictions for Event-B machines., Comment: 22 pages

Abstract State Machines (ASMs) provide a model of computations on structures rather than strings. Blass, Gurevich and Shelah showed that deterministic PTIME-bounded ASMs define the choiceless fragment of PTIME, but cannot capture PTIME. In this article deterministic PSPACE-bounded ASMs are introduced, and it is proven that they cannot capture PSPACE. The key for the proof is a characterisation by partial fixed-point formulae over the St\"ark/Nanchen logic for deterministic ASMs and a construction of transitive structures, in which such formulae must hold. This construction exploits that the decisive support theorem for choiceless polynomial time holds under slightly weaker assumptions., Comment: 12 pages. arXiv admin note: substantial text overlap with arXiv:2005.04598

Regular decision processes (RDPs) are a subclass of non-Markovian decision processes where the transition and reward functions are guarded by some regular property of the past (a lookback). While RDPs enable intuitive and succinct representation of non-Markovian decision processes, their expressive power coincides with finite-state Markov decision processes (MDPs). We introduce omega-regular decision processes (ODPs) where the non-Markovian aspect of the transition and reward functions are extended to an omega-regular lookahead over the system evolution. Semantically, these lookaheads can be considered as promises made by the decision maker or the learning agent about her future behavior. In particular, we assume that, if the promised lookaheads are not met, then the payoff to the decision maker is $\bot$ (least desirable payoff), overriding any rewards collected by the decision maker. We enable optimization and learning for ODPs under the discounted-reward objective by reducing them to lexicographic optimization and learning over finite MDPs. We present experimental results demonstrating the effectiveness of the proposed reduction.

While discounted payoff games and classic games that reduce to them, like parity and mean-payoff games, are symmetric, their solutions are not. We have taken a fresh view on the constraints that optimal solutions need to satisfy, and devised a novel way to converge to them, which is entirely symmetric. It also challenges the gospel that methods for solving payoff games are either based on strategy improvement or on value iteration., Comment: In Proceedings GandALF 2023, arXiv:2309.17318

Reinforcement learning (RL) is a powerful approach for training agents to perform tasks, but designing an appropriate reward mechanism is critical to its success. However, in many cases, the complexity of the learning objectives goes beyond the capabilities of the Markovian assumption, necessitating a more sophisticated reward mechanism. Reward machines and omega-regular languages are two formalisms used to express non-Markovian rewards for quantitative and qualitative objectives, respectively. This paper introduces omega-regular reward machines, which integrate reward machines with omega-regular languages to enable an expressive and effective reward mechanism for RL. We present a model-free RL algorithm to compute epsilon-optimal strategies against omega-egular reward machines and evaluate the effectiveness of the proposed algorithm through experiments., Comment: To appear in ECAI-2023

Good-for-MDPs and good-for-games automata are two recent classes of nondeterministic automata that reside between general nondeterministic and deterministic automata. Deterministic automata are good-for-games, and good-for-games automata are good-for-MDPs, but not vice versa. One of the question this raises is how these classes relate in terms of succinctness. Good-for-games automata are known to be exponentially more succinct than deterministic automata, but the gap between good-for-MDPs and good-for-games automata as well as the gap between ordinary nondeterministic automata and those that are good-for-MDPs have been open. We establish that these gaps are exponential, and sharpen this result by showing that the latter gap remains exponential when restricting the nondeterministic automata to separating safety or unambiguous reachability automata., Comment: 18 pages

Families of DFAs (FDFAs) have recently been introduced as a new representation of $\omega$-regular languages. They target ultimately periodic words, with acceptors revolving around accepting some representation $u\cdot v^\omega$. Three canonical FDFAs have been suggested, called periodic, syntactic, and recurrent. We propose a fourth one, limit FDFAs, which can be exponentially coarser than periodic FDFAs and are more succinct than syntactic FDFAs, while they are incomparable (and dual to) recurrent FDFAs. We show that limit FDFAs can be easily used to check not only whether {\omega}-languages are regular, but also whether they are accepted by deterministic B\"uchi automata. We also show that canonical forms can be left behind in applications: the limit and recurrent FDFAs can complement each other nicely, and it may be a good way forward to use a combination of both. Using this observation as a starting point, we explore making more efficient use of Myhill-Nerode's right congruences in aggressively increasing the number of don't-care cases in order to obtain smaller progress automata. In pursuit of this goal, we gain succinctness, but pay a high price by losing constructiveness., Comment: 30 pages, accepted at ATVA 2023

We introduce a method for translating an alternating weak B\"uchi automaton (AWA), which corresponds to a Linear Dynamic Logic (LDL) formula, to an unambiguous B\"uchi automaton (UBA). Our translations generalise constructions for Linear Temporal Logic (LTL), a less expressive specification language than LDL. In classical constructions, LTL formulas are first translated to alternating \emph{very weak} automata (AVAs) -- automata that have only singleton strongly connected components (SCCs); the AVAs are then handled by efficient disambiguation procedures. However, general AWAs can have larger SCCs, which complicates disambiguation. Currently, the only available disambiguation procedure has to go through an intermediate construction of nondeterministic B\"uchi automata (NBAs), which would incur an exponential blow-up of its own. We introduce a translation from \emph{general} AWAs to UBAs with a \emph{singly} exponential blow-up, which also immediately provides a singly exponential translation from LDL to UBAs. Interestingly, the complexity of our translation is \emph{smaller} than the best known disambiguation algorithm for NBAs (broadly $(0.53n)^n$ vs. $(0.76n)^n$), while the input of our construction can be exponentially more succinct., Comment: 23 pages

We explore the notion of history-determinism in the context of timed automata (TA) over infinite timed words. History-deterministic (HD) automata are those in which nondeterminism can be resolved on the fly, based on the run constructed thus far. History-determinism is a robust property that admits different game-based characterisations, and HD specifications allow for game-based verification without an expensive determinization step. We show that the class of timed $\omega$-languages recognized by HD timed automata strictly extends that of deterministic ones, and is strictly included in those recognised by fully non-deterministic TA. For non-deterministic timed automata it is known that universality is already undecidable for safety/reachability TA. For history-deterministic TA with arbitrary parity acceptance, we show that timed universality, inclusion, and synthesis all remain decidable and are EXPTIME-complete. For the subclass of TA with safety or reachability acceptance, one can decide (in EXPTIME) whether such an automaton is history-deterministic. If so, it can effectively determinized without introducing new automaton states.

Cryogenic buffer gas beams are central to many cold molecule experiments. Here, we use absorption and fluorescence spectroscopy to directly compare molecular beams of AlF, CaF, MgF, and YbF molecules, produced by chemical reaction of laser ablated atoms with fluorine rich reagents. The beam brightness for AlF is measured as $2\times 10^{12}$ molecules per steradian per pulse in a single rotational state, comparable to an Al atomic beam produced in the same setup. The CaF, MgF and YbF beams show an order of magnitude lower brightness than AlF, and far below the brightness of Ca and Yb beams. The addition of either NF$_3$ or SF$_6$ to the cell extinguishes the Al atomic beam, but has a minimal effect on the Ca and Yb beams. NF$_3$ reacts more efficiently than SF$_6$, as a significantly lower flow rate is required to maximise the molecule production, which is particularly beneficial for long-term stability of the AlF beam. We use NO as a proxy for the reactant gas as it can be optically detected. We demonstrate that a cold, rotationally pure NO beam can be generated by laser desorption, thereby gaining insight into the dynamics of the reactant gas inside the buffer gas cell., Comment: 29 pages, 17 figures, 1 table

Birch reduction is a time-proven way to hydrogenate aromatic hydrocarbons (such as benzene), which relies on the reducing power of electrons released from alkali metals into liquid ammonia. We have succeeded to characterize the key intermediates of the Birch reduction process - the solvated electron and dielectron and the benzene radical anion - using cyclic voltammetry and photoelectron spectroscopy, aided by electronic structure calculations. In this way, we not only quantify the electron binding energies of these species, which are decisive for the mechanism of the reaction but also use Birch reduction as a case study to directly connect the two seemingly unrelated experimental techniques., Comment: after revision

While finite automata have minimal DFAs as a simple and natural normal form, deterministic omega-automata do not currently have anything similar. One reason for this is that a normal form for omega-regular languages has to speak about more than acceptance - for example, to have a normal form for a parity language, it should relate every infinite word to some natural color for this language. This raises the question of whether or not a concept such as a natural color of an infinite word (for a given language) exists, and, if it does, how it relates back to automata. We define the natural color of a word purely based on an omega-regular language, and show how this natural color can be traced back from any deterministic parity automaton after two cheap and simple automaton transformations. The resulting streamlined automaton does not necessarily accept every word with its natural color, but it has a 'co-run', which is like a run, but can once move to a language equivalent state, whose color is the natural color, and no co-run with a higher color exists. The streamlined automaton defines, for every color c, a good-for-games co-B\"uchi automaton that recognizes the words whose natural colors w.r.t. the represented language are at least c. This provides a canonical representation for every $\omega$-regular language, because good-for-games co-B\"uchi automata have a canonical minimal (and cheap to obtain) representation for every co-B\"uchi language.

Recursion is the fundamental paradigm to finitely describe potentially infinite objects. As state-of-the-art reinforcement learning (RL) algorithms cannot directly reason about recursion, they must rely on the practitioner's ingenuity in designing a suitable "flat" representation of the environment. The resulting manual feature constructions and approximations are cumbersome and error-prone; their lack of transparency hampers scalability. To overcome these challenges, we develop RL algorithms capable of computing optimal policies in environments described as a collection of Markov decision processes (MDPs) that can recursively invoke one another. Each constituent MDP is characterized by several entry and exit points that correspond to input and output values of these invocations. These recursive MDPs (or RMDPs) are expressively equivalent to probabilistic pushdown systems (with call-stack playing the role of the pushdown stack), and can model probabilistic programs with recursive procedural calls. We introduce Recursive Q-learning -- a model-free RL algorithm for RMDPs -- and prove that it converges for finite, single-exit and deterministic multi-exit RMDPs under mild assumptions.

When omega-regular objectives were first proposed in model-free reinforcement learning (RL) for controlling MDPs, deterministic Rabin automata were used in an attempt to provide a direct translation from their transitions to scalar values. While these translations failed, it has turned out that it is possible to repair them by using good-for-MDPs (GFM) B\"uchi automata instead. These are nondeterministic B\"uchi automata with a restricted type of nondeterminism, albeit not as restricted as in good-for-games automata. Indeed, deterministic Rabin automata have a pretty straightforward translation to such GFM automata, which is bi-linear in the number of states and pairs. Interestingly, the same cannot be said for deterministic Streett automata: a translation to nondeterministic Rabin or B\"uchi automata comes at an exponential cost, even without requiring the target automaton to be good-for-MDPs. Do we have to pay more than that to obtain a good-for-MDP automaton? The surprising answer is that we have to pay significantly less when we instead expand the good-for-MDP property to alternating automata: like the nondeterministic GFM automata obtained from deterministic Rabin automata, the alternating good-for-MDP automata we produce from deterministic Streett automata are bi-linear in the the size of the deterministic automaton and its index, and can therefore be exponentially more succinct than minimal nondeterministic B\"uchi automata.

Calude et al. have recently shown that parity games can be solved in quasi-polynomial time, a landmark result that has led to a number of approaches with quasi-polynomial complexity. Jurdinski and Lasic have further improved the precise complexity of parity games, especially when the number of priorities is low (logarithmic in the number of positions). Both of these algorithms belong to a class of game solving techniques now often called separating automata: deterministic automata that can be used as witness automata to decide the winner in parity games up to a given number of states and colours. We suggest a number of adjustments to the approach of Calude et al. that lead to smaller statespaces. These include and improve over those earlier introduced by Fearnley et al. We identify two of them that, together, lead to a statespace of exactly the same size Jurdzinski and Lasic's concise progress measures, which currently hold the crown as smallest statespace. The remaining improvements, hence, lead to a further reduction in the size of the statespace, making our approach the most succinct progress measures available for parity games.

Adversarial training has been shown to be one of the most effective approaches to improve the robustness of deep neural networks. It is formalized as a min-max optimization over model weights and adversarial perturbations, where the weights can be optimized through gradient descent methods like SGD. In this paper, we show that treating model weights as random variables allows for enhancing adversarial training through \textbf{S}econd-Order \textbf{S}tatistics \textbf{O}ptimization (S$^2$O) with respect to the weights. By relaxing a common (but unrealistic) assumption of previous PAC-Bayesian frameworks that all weights are statistically independent, we derive an improved PAC-Bayesian adversarial generalization bound, which suggests that optimizing second-order statistics of weights can effectively tighten the bound. In addition to this theoretical insight, we conduct an extensive set of experiments, which show that S$^2$O not only improves the robustness and generalization of the trained neural networks when used in isolation, but also integrates easily in state-of-the-art adversarial training techniques like TRADES, AWP, MART, and AVMixup, leading to a measurable improvement of these techniques. The code is available at \url{https://github.com/Alexkael/S2O}., Comment: Accepted by CVPR2022

Aluminum monofluoride (AlF) possesses highly favorable properties for laser cooling, both via the A$^1\Pi$ and a$^3\Pi$ states. Determining efficient pathways between the singlet and the triplet manifold of electronic states will be advantageous for future experiments at ultralow temperatures. The lowest rotational levels of the A$^1\Pi, v=6$ and b$^3\Sigma^+, v=5$ states of AlF are nearly iso-energetic and interact via spin-orbit coupling. These levels thus have a strongly mixed spin-character and provide a singlet-triplet doorway. We here present a hyperfine resolved spectroscopic study of the A$^1\Pi, v=6$ // b$^3\Sigma^+, v=5$ perturbed system in a jet-cooled, pulsed molecular beam. From a fit to the observed energies of the hyperfine levels, the fine and hyperfine structure parameters of the coupled states, their relative energies as well as the spin-orbit interaction parameter are determined. The standard deviation of the fit is about 15 MHz. We experimentally determine the radiative lifetimes of selected hyperfine levels by time-delayed ionization, Lamb dip spectroscopy and accurate measurements of the transition lineshapes. The measured lifetimes range between 2 ns and 200 ns, determined by the degree of singlet-triplet mixing for each level., Comment: 12 pages, 11 figures

Nondeterministic Good-for-MDP (GFM) automata are for MDP model checking and reinforcement learning what good-for-games automata are for reactive synthesis: a more compact alternative to deterministic automata that displays nondeterminism, but only so much that it can be resolved locally, such that a syntactic product can be analysed. GFM has recently been introduced as a property for reinforcement learning, where the simpler B\"uchi acceptance conditions it allows to use is key. However, while there are classic and novel techniques to obtain automata that are GFM, there has not been a decision procedure for checking whether or not an automaton is GFM. We show that GFM-ness is decidable and provide an EXPTIME decision procedure as well as a PSPACE-hardness proof., Comment: 26 pages, accepted to CONCUR 2023

While dropout is known to be a successful regularization technique, insights into the mechanisms that lead to this success are still lacking. We introduce the concept of \emph{weight expansion}, an increase in the signed volume of a parallelotope spanned by the column or row vectors of the weight covariance matrix, and show that weight expansion is an effective means of increasing the generalization in a PAC-Bayesian setting. We provide a theoretical argument that dropout leads to weight expansion and extensive empirical support for the correlation between dropout and weight expansion. To support our hypothesis that weight expansion can be regarded as an \emph{indicator} of the enhanced generalization capability endowed by dropout, and not just as a mere by-product, we have studied other methods that achieve weight expansion (resp.\ contraction), and found that they generally lead to an increased (resp.\ decreased) generalization ability. This suggests that dropout is an attractive regularizer, because it is a computationally cheap method for obtaining weight expansion. This insight justifies the role of dropout as a regularizer, while paving the way for identifying regularizers that promise improved generalization through weight expansion., Comment: TMLR

We present chemical kinetics measurements of the luminol oxydation chemiluminescence reaction at the interface between two aqueous solutions, using liquid jet technology. Free-flowing iquid microjets are a relatively recent development that has found its way into a growing number of applications in spectroscopy and dynamics. A variant thereof, called flat-jet, is obtained when two cylindrical jets of a liquid are crossed, leading to a chain of planar leaf-shaped structures of the flowing liquid. We here show that in the first leaf of this chain the fluids do not exhibit turbulent mixing, providing a clean interface between the liquids from the impinging jets. We also show, using the example of the luminol chemiluminescence reaction, how this setup can be used to obtain kinetics information from friction-less flow and by circumventing the requirement for rapid mixing but by intentionally suppressing all turbulent mixing and instead relying on diffusion.

The increasing use of Machine Learning (ML) components embedded in autonomous systems -- so-called Learning-Enabled Systems (LESs) -- has resulted in the pressing need to assure their functional safety. As for traditional functional safety, the emerging consensus within both, industry and academia, is to use assurance cases for this purpose. Typically assurance cases support claims of reliability in support of safety, and can be viewed as a structured way of organising arguments and evidence generated from safety analysis and reliability modelling activities. While such assurance activities are traditionally guided by consensus-based standards developed from vast engineering experience, LESs pose new challenges in safety-critical application due to the characteristics and design of ML models. In this article, we first present an overall assurance framework for LESs with an emphasis on quantitative aspects, e.g., breaking down system-level safety targets to component-level requirements and supporting claims stated in reliability metrics. We then introduce a novel model-agnostic Reliability Assessment Model (RAM) for ML classifiers that utilises the operational profile and robustness verification evidence. We discuss the model assumptions and the inherent challenges of assessing ML reliability uncovered by our RAM and propose solutions to practical use. Probabilistic safety argument templates at the lower ML component-level are also developed based on the RAM. Finally, to evaluate and demonstrate our methods, we not only conduct experiments on synthetic/benchmark datasets but also scope our methods with case studies on simulated Autonomous Underwater Vehicles and physical Unmanned Ground Vehicles., Comment: Preprint Accepted by ACM Transactions on Embedded Computing Systems

This work studies equilibrium problems under uncertainty where firms maximize their profits in a robust way when selling their output. Robust optimization plays an increasingly important role when best guaranteed objective values are to be determined, independently of the specific distributional assumptions regarding uncertainty. In particular, solutions are to be determined that are feasible regardless of how the uncertainty manifests itself within some predefined uncertainty set. Our mathematical analysis adopts the robust optimization perspective in the context of equilibrium problems. First, we present structural insights for a single-stage, nonadjustable robust setting. We then go one step further and study the more complex two-stage or adjustable case where a part of the variables can adjust to the realization of the uncertainty. We compare equilibrium outcomes with the corresponding centralized robust optimization problem where thesum of all profits are maximized. As we find, the market equilibrium for the perfectly competitive firms differs from the solution of the robust central planner, which is in stark contrast to classical results regarding the efficiency of market equilibria with perfectly competitive firms. For the different scenarios considered, we furthermore are able to determine the resulting price of anarchy. In the case of non-adjustable robustness, for fixed demand in every time step the price of anarchy is bounded whereas it is unbounded if the buyers are modeled by elastic demand functions. For the two-stage adjustable setting, we show how to compute subsidies for the firms that lead to robust welfareoptimal equilibria., Comment: 26 pages