Feature-based scenario clustering and selective degrees of freedom reduction for two-stage stochastic optimization of Power-to-Methanol

This paper proposes a two-stage optimization approach for aerocapture guidance. In classical entry guidance systems, deterministic optimization is used. Large-scale and short-scale density perturbations may strongly affect the performance of the guidance system, and variations in those are usually not accounted for when computing the command. In this work, perturbations that affect the trajectory at future time-steps are taken into consideration when computing the commanded bank angle. The chosen command is optimal based on a set of possible future perturbations, after the observation of which, a correction can be made. Both two-stage stochastic and two-stage robust optimization are proposed as a solution. In a Monte Carlo analysis consisting of 50 runs, the two-stage robust optimization guidance outperforms an optimal, deterministic, numeric predictor-corrector guidance. Excluding one outlier, also the two-stage stochastic optimization makes the guidance perform better than an optimal deterministic numeric predictor-corrector. With either approach, the computational demands are increased by about thirty times compared to an optimal numeric predictor-corrector. Much of the computation time increase may be reduced by parallelization. On the other hand, the extensive tuning required for the optimal numeric predictor-corrector is not needed for the two-stage optimization guidance, making this approach conceptually more robust. Better modeling of the environment may help further improve the performance. Finally, an approximation to the two-stage robust optimization approach is developed. The guidance has computational requirements only four times larger than those of the optimal numeric predictor-corrector guidance, but can be parallelized into two threads, and, except for two outliers, it offers improved performance.

Reliability-informed end-of-use decision making for product sustainability using two-stage stochastic optimization

Two-stage stochastic optimization, in which a long-term master problem is coupled with a family of short-term subproblems, plays a critical role in various application areas. However, most existing algorithms for two-stage stochastic optimization only work for special cases, and/or are based on the batch method, which requires huge memory and computational complexity. To the best of our knowledge, there still lack efficient and general two-stage online stochastic optimization algorithms. This paper proposes a two-stage online successive convex approximation (TOSCA) algorithm for general two-stage nonconvex stochastic optimization problems. At each iteration, the TOSCA algorithm first solves one short-term subproblem associated with the current realization of the system state. Then, it constructs a convex surrogate function for the objective of the long-term master problem. Finally, the long-term variables are updated by solving a convex approximation problem obtained by replacing the objective function in the long-term master problem with the convex surrogate function. We establish the almost sure convergence of the TOSCA algorithm and customize the algorithmic framework to solve three important application problems. Simulations show that the TOSCA algorithm can achieve superior performance over existing solutions.

We describe a two-stage robust optimization approach for solving network flow and design problems with uncertain demand. In two-stage network optimization, one defers a subset of the flow decisions until after the realization of the uncertain demand. Availability of such a recourse action allows one to come up with less conservative solutions compared to single-stage optimization. However, this advantage often comes at a price: two-stage optimization is, in general, significantly harder than single-stage optimization. For network flow and design under demand uncertainty, we give a characterization of the first-stage robust decisions with an exponential number of constraints and prove that the corresponding separation problem is 𝒩𝒫-hard even for a network flow problem on a bipartite graph. We show, however, that if the second-stage network topology is totally ordered or an arborescence, then the separation problem is tractable. Unlike single-stage robust optimization under demand uncertainty, two-stage robust optimization allows one to control conservatism of the solutions by means of an allowed “budget for demand uncertainty.” Using a budget of uncertainty, we provide an upper bound on the probability of infeasibility of a robust solution for a random demand vector. We generalize the approach to multicommodity network flow and design, and give applications to lot-sizing and location-transportation problems. By projecting out second-stage flow variables, we define an upper bounding problem for the two-stage min-max-min optimization problem. Finally, we present computational results comparing the proposed two-stage robust optimization approach with single-stage robust optimization as well as scenario-based two-stage stochastic optimization.

A new scheme to cope with two-stage stochastic optimization problems uses a risk measure as the objective function of the recourse action, where the risk measure is defined as the worst-case expected values over a set of constrained distributions. This paper develops an approach to deal with the case where both the first and second stage objective functions are convex linear-quadratic. It is shown that under a standard set of regularity assumptions, this two-stage quadratic stochastic optimization problem with measures of risk is equivalent to a conic optimization problem that can be solved in polynomial time.

Two-stage stochastic optimization frameworks to aid in decision-making under uncertainty for variable resource generators participating in a sequential energy market

We present two-stage stochastic risk averse optimization models for the power generation mix capacity expansion planning in the long run under uncertainty. Uncertainty is described by a set of possible scenarios in the second stage and uncertain parameters are the unit production costs of the existing power plants as well as those of the candidate plants of new technologies among which to choose, the market electricity price, the price of green certificates and the emission permits and the potential market share of the producer. The problem is expressed as a two-stage stochastic integer optimization model subject to technical constraints, market opportunities and budget constraints. First stage variables represent the number of new power plants for each candidate technology to be added to the existing generation mix every year of the planning horizon. Second stage variables are the continuous operation variables of all power plants in the generation mix along the time horizon. We solve the problem of the maximization of the net present value of the expected profits along the time horizon using both a risk neutral approach and different risk averse strategies (conditional value at risk, shortfall probability, expected shortage and first- and second-order stochastic dominance), under different hypotheses of the available budget, analysing the impact of each risk averse strategy on the expected profit. Results show that risk control strongly reduces the possibility of reaching unwanted scenarios as well as providing consistent solutions under different strategies.

During the periods of high heat load (HL) demand in winter, the increased heat output of combined heat and power units (CHPs) can significantly compress the electric power regulation range, thereby posing a threat to the safety of the power system. The solid electric thermal storage (SETS) can be employed as the regulating resource for both electric and thermal systems, expanding the dispatch space of microgrids to promote renewable energy consumption. In this paper, a two-stage stochastic robust optimization scheduling model of an electric–thermal microgrid with SETS is proposed, and the electric–thermal bi-directional regulation characteristics of SETS are considered. First, a SETS operation model based on the heat transfer characteristics of the homogeneous material-type thermal storage unit is established. Second, a regional thermal inertia model under a constant-flow variable-temperature system is established, which integrates HL at different locations to the heat source, avoiding the calculation of variation in water temperature, thus reducing the calculation difficulty. Finally, the two-stage stochastic robust optimization scheduling model of an electric–thermal microgrid with SETS is established. The model decouples the power and heat generation of CHP through the bi-directional regulation function of SETS. Case studies demonstrate the validity and effectiveness of the proposed scheduling model.

A new two-stage stochastic partial differential equation (PDE)-constrained optimization methodology is developed for the active vibration control of structures in the presence of uncertainties in mechanical loads. The methodology relies on the two-stage stochastic optimization formulation with an embedded first-order black-box PDE-constrained optimization procedure. The PDE-constrained optimization procedure utilizes a first-order active-set algorithm with a conjugate gradient method. The objective function is determined through solution of the governing PDEs and its gradient is computed using automatic differentiation with hyper-dual numbers. The developed optimization methodology is applied to the problem of post-impact vibration control (via applied electromagnetic field) of an electrically conductive carbon fiber reinforced composite plate subjected to an uncertain, or stochastic, impact load. The corresponding governing PDEs consist of a nonlinear coupled system of equations of motion and Maxwell's equations. The conducted computational study shows that the obtained two-stage optimization solution allows for a significant suppression of vibrations caused by the randomized impact load in all impact load scenarios. Also, the effectiveness of the developed methodology is illustrated in the case of a deterministic impact load, where the two-stage strategy enables one to practically eliminate post-impact vibrations.

Following the well-studied two-stage optimization framework for stochastic optimization [15,8], we study approximation algorithms for robust two-stage optimization problems with an exponential number of scenarios. Prior to this work, Dhamdhere et al. [8] introduced approximation algorithms for two-stage robust optimization problems with explicitly given scenarios. In this paper, we assume the set of possible scenarios is given implicitly, for example by an upper bound on the number of active clients. In two-stage robust optimization, we need to pre-purchase some resources in the first stage before the adversary's action. In the second stage, after the adversary chooses the clients that need to be covered, we need to complement our solution by purchasing additional resources at an inflated price. The goal is to minimize the cost in the worst-case scenario. We give a general approach for solving such problems using LP rounding. Our approach uncovers an interesting connection between robust optimization and online competitive algorithms. We use this approach, together with known online algorithms, to develop approximation algorithms for several robust covering problems, such as set cover, vertex cover, and edge cover. We also study a simple buy-at-oncealgorithm that either covers all items in the first stage or does nothing in the first stage and waits to build the complete solution in the second stage. We show that this algorithm gives tight approximation factors for unweighted variants of these covering problems, but performs poorly for general weighted problems.

Surgery demand is an uncertain parameter in addressing the problem of surgery block allocations, and its typical variability should be considered to ensure the feasibility of surgical planning. We develop two models, a stochastic recourse programming model and a two-stage stochastic optimization (SO) model with incorporated risk measure terms in the objective functions to determine a planning decision that is made to allocate surgical specialties to operating rooms (ORs). Our aim is to minimize the costs associated with postponements and unscheduled demands as well as the inefficient use of OR capacity. The results of these models are compared using a case of a real-life hospital to determine which model better copes with uncertainty. We propose a novel framework to transform the SO model based on its deterministic counterpart. Three SO models are proposed with respect to the variability and infeasibility of the measures of the objective function to encode the construction of the SO framework. The analysis of the experimental results demonstrates that the SO model offers better performance under a highly volatile demand environment than the recourse model. The originality of this work lies in its use of SO transformation framework and its development of stochastic models to address the problem of surgery capacity allocation based on a real case.

The emerging distributed satellite cluster network (DSCN) holds great promise in various practical fields, including earth observation, disaster rescue, and tracking of forest fires. In the DSCN environment, it is essential to achieve the best data delivery performance by coordinating multi-dimensional heterogeneous and dynamic resources. However, in real-world applications, the distribution of long-term data arrival is not often fully known. Motivated by this fact, we propose a distributionally robust two-stage stochastic optimization framework with considering the dynamic network resources and the partially known distribution information of long-term data arrival. Aiming at maximizing the total network reward, we formulate a two-stage stochastic flow optimization problem based on the extended time expanded graph. Then, we introduce an ambiguity set for the uncertain distribution of the long-term random data arrival inspired by the idea from the distributionally robust optimization. On the basis of the proposed ambiguity set, we further propose a data arrival distribution robust two-stage recourse (DADR-TR) algorithm by converting the original stochastic optimization problem into a deterministic cone optimization problem, which is computationally tractable. The extensive simulations have been conducted to evaluate the impact of various network parameters on the algorithm performance and further validate that the proposed DADR-TR algorithm can achieve high data delivery performance without full distribution information of the long-term data arrival.

In this paper, we update the theory of deterministic linear semi-infinite programming, mainly with the dual characterizations of the constraint qualifications, which play a crucial role in optimality and duality. From this theory, we obtain new results on conic and two-stage stochastic linear optimization. Specifically, for conic linear optimization problems, we characterize the existence of feasible solutions and some geometric properties of the feasible set, and we also provide theorems on optimality and duality. Analogously, regarding stochastic optimization problems, we study the semi-infinite reformulation of a problem-based scenario reduction problem in two-stage stochastic linear programming, providing a sufficient condition for the existence of feasible solutions as well as optimality and duality theorems to its non-combinatorial part.

A scenario-based two-stage stochastic optimization approach for multi-energy microgrids

With renewable energy sources (RESs) highly penetrating into the power system, new problems emerge for the independent system operator (ISO) to maintain and keep the power system safe and reliable in the day-ahead dispatching process under the fluctuation caused by renewable energy. In this paper, considering the small hydropower with no reservoir, different from the other hydro optimization research and wind power uncertain circumstances, a day-ahead scheduling model is proposed for a distributed power grid system which contains several distributed generators, such as small hydropower and wind power, and energy storage systems. To solve this model, a two-stage stochastic robust optimization approach is presented to smooth out hydro power and wind power output fluctuation with the aim of minimizing the total expected system operation cost under multiple cluster water inflow scenarios, and the worst case of wind power output uncertainty. More specifically, before dispatching and clearing, it is necessary to cluster the historical inflow scenarios of small hydropower into several typical scenarios via the Fuzzy C-means (FCM) clustering method, and then the clustering comprehensive quality (CCQ) method is also presented to evaluate whether these scenarios are representative, which has previously been ignored by cluster research. It can be found through numerical examples that FCM-CCQ can explain the classification more reasonably than the common clustering method. Then we optimize the two stage scheduling, which contain the pre-clearing stage and the rescheduling stage under each typical inflow scenario after clustering, and then calculate the final operating cost under the worst wind power output scenario. To conduct the proposed model, the day-ahead scheduling procedure on the Institute of Electrical and Electronics Engineers (IEEE) 30-bus test system is simulated with real hydropower and wind power data. Compared with traditional deterministic optimization, the results of two-stage stochastic robust optimization structured in this paper, increases the total cost of the system, but enhances the conservative scheduling strategy, improves the stability and reliability of the power system, and reduces the risk of decision-making simultaneously.