An efficient linear model and optimisation algorithm for multi-site combined heat and power production

Combined heat and power (CHP) production is an important energy production technology which can help to improve the efficiency of energy production and to reduce the emission of CO₂. Cost-efficient operation of a CHP system can be planned using an optimisation model based on hourly load forecasts. A long-term planning model decomposes into hourly models, which can be formulated as linear programming (LP) problems. Such problems can be solved efficiently using the specialized Power Simplex algorithm. However, Power Simplex can only manage one heat and one power balance. Since heat cannot be transported over long distances, Power Simplex applies only for local CHP planning.In this paper we formulate the hourly multi-site CHP planning problem with multiple heat balances as an LP model with a special structure. We then develop the Extended Power Simplex (EPS) algorithm for solving such models efficiently. Even though the problem can be quite large as the number of demand sites increases, EPS demonstrates very good efficiency. In test runs with realistic models, EPS is from 29 to 85 times faster than an efficient sparse Simplex code using the product form of inverse (PFI). Furthermore, the relative efficiency of EPS improves as the problem size grows.     

A variant of the dynamic programming algorithm for unit commitment of combined heat and power systems

The paper addresses the unit commitment in multi-period combined heat and power (CHP) production planning under the deregulated power market. In CHP plants (units), generation of heat and power follows joint characteristics, which means that production planning must be done in coordination. We introduce in this paper the DP-RSC1 algorithm, which is a variant of the dynamic programming (DP) algorithm based on linear relaxation of the ON/OFF states of the units and sequential commitment of units one by one. The time complexity of DP-RSC1 is proportional to the number of generating units in the system, the number of periods over the planning horizon and the time for solving a single-period economic dispatch problem. We have compared the DP-RSC1 algorithm with realistic power plants against the unit decommitment algorithm and the traditional priority listing method. The results show that the DP-RSC1 algorithm gives somewhat more accurate results (0.08–0.5% on average, maximum 10% for the individual sub-case) and executes 3–5 times faster on average than the unit decommitment algorithm. It is not surprising that the solution quality of the DP-RSC1 algorithm is much better than that of the priority listing method.     

An improved unit decommitment algorithm for combined heat and power systems

This paper addresses the unit commitment in multi-period combined heat and power (CHP) production planning, considering the possibility to trade power on the spot market. In CHP plants (units), generation of heat and power follows joint characteristics, which means that production planning for both heat and power must be done in coordination. We present an improved unit decommitment (IUD) algorithm that starts with an improved initial solution with less heat surplus so that the relative cost-efficiency of the plants can be determined more accurately. Then the subsequent decommitment procedures can decommit (switch off) the least cost-efficient plants properly. The improved initial solution for the committed plants is generated by a heuristic procedure. The heuristic procedure utilizes both the Lagrangian relaxation principle that relaxes the system-wide (heat and power) demand constraints and a linear relaxation of the ON/OFF states of the plants. We compare the IUD algorithm with realistic test data against a generic unit decommitment (UD) algorithm. Numerical results show that IUD is an overall improvement of UD. The solution quality of IUD is better than that of UD for almost all of tested cases. The maximum improvement is 11.3% and the maximum degradation is only 0.04% (only two sub-cases out of 216 sub-cases) with an average improvement of 0.3–0.5% for different planning horizons. Moreover, IUD is more efficient (1.1–3 times faster on average) than UD.     

An efficient envelope-based Branch and Bound algorithm for non-convex combined heat and power production planning

Combined heat and power (CHP) production is universally accepted as one of the most energy-efficient technologies to produce energy with lower fuel consumption and fewer emissions. In CHP technology, heat and power generation follow a joint characteristic. Traditional CHP production is usually applied in backpressure plants, where the joint characteristic can often be represented by a convex model. Advanced CHP production technologies such as backpressure plants with condensing and auxiliary cooling options, gas turbines, and combined gas and steam cycles may require non-convex models. Cost-efficient operation of a CHP system can be planned using an optimization model based on forecasts for heat load and power price. A long-term planning model decomposes into thousands of single-period models, which can be formulated in the convex case as linear programming (LP) problems, and in the non-convex case as mixed integer programming (MIP) problems.     

Non-convex power plant modelling in energy optimisation

The European electricity market has been deregulated recently. This means that energy companies must optimise power generation considering the rapidly fluctuating price on the spot market. Optimisation has also become more difficult. New production technologies, such as gas turbines (GT), combined heat and power generation (CHP), and combined steam and gas cycles (CSG) require non-convex models. Risk analysis through stochastic simulation requires solving a large number of models rapidly. These factors have created a need for more versatile and efficient decision-support tools for energy companies.We formulate the decision-problem of a power company as a large mixed integer programming (MIP) model. To make the model manageable we compose the model hierarchically from modular components. To speed up the optimisation procedure, we decompose the problem into hourly sub-problems, and develop a customised Branch-and-Bound algorithm for solving the sub-problems efficiently. We demonstrate the use of the model with a real-life application.     

Lagrangian relaxation based algorithm for trigeneration planning with storages

Trigeneration is a booming power production technology where three energy commodities are simultaneously produced in a single integrated process. Electric power, heat (e.g. hot water) and cooling (e.g. chilled water) are three typical energy commodities in the trigeneration system. The production of three energy commodities follows a joint characteristic. This paper presents a Lagrangian relaxation (LR) based algorithm for trigeneration planning with storages based on deflected subgradient optimization method. The trigeneration planning problem is modeled as a linear programming (LP) problem. The linear cost function poses the convergence challenge to the LR algorithm and the joint characteristic of trigeneration plants makes the operating region of trigeneration system more complicated than that of power-only generation system and that of combined heat and power (CHP) system. We develop an effective method for the long-term planning problem based on the proper strategy to form Lagrangian subproblems and solve the Lagrangian dual (LD) problem based on deflected subgradient optimization method. We also develop a heuristic for restoring feasibility from the LD solution. Numerical results based on realistic production models show that the algorithm is efficient and near-optimal solutions are obtained.     

A specialized network simplex algorithm for the constrained maximum flow problem

The constrained maximum flow problem is to send the maximum flow from a source to a sink in a directed capacitated network where each arc has a cost and the total cost of the flow cannot exceed a budget. This problem is similar to some variants of classical problems such as the constrained shortest path problem, constrained transportation problem, or constrained assignment problem, all of which have important applications in practice. The constrained maximum flow problem itself has important applications, such as in logistics, telecommunications and computer networks. In this research, we present an efficient specialized network simplex algorithm that significantly outperforms the two widely used LP solvers: CPLEX and lp_solve. We report CPU times of an average of 27 times faster than CPLEX (with its dual simplex algorithm), the closest competitor of our algorithm.     

Joint planning of combined heat and power and electric power systems: An efficient model formulation

The growing importance of combined heat and power (CHP) around the world has increased the need to consider its role within electric power systems. In this paper, we show how the problem of joint planning of CHP and electric power systems may be formulated more efficiently than has previously been done, by exploiting the special structure of the optimal solution. Numerical tests indicate that this reformulation typically allows a reduction in the time required for problem solution by a factor of two to five.     

CO2 emissions trading planning in combined heat and power production via multi-period stochastic optimization

The EU emissions trading scheme (ETS) taking effect in 2005 covers CO₂ emissions from specific large-scale industrial activities and combustion installations. A large number of existing and potential future combined heat and power (CHP) installations are subject to ETS and targeted for emissions reduction. CHP production is an important technology for efficient and clean provision of energy because of its superior carbon efficiency. The proper planning of emissions trading can help its potential into full play, making it become a true “winning technology” under ETS. Fuel mix or fuel switch will be the reasonable choices for fossil fuel based CHP producers to achieve their emissions targets at the lowest possible cost. In this paper we formulate CO₂ emissions trading planning of a CHP producer as a multi-period stochastic optimization problem and propose a stochastic simulation and coordination approach for considering the risk attitude of the producer, penalty for excessive emissions, and the confidence interval for emission estimates. In test runs with a realistic CHP production model, the proposed solution approach demonstrates good trading efficiency in terms of profit-to-turnover ratio. Considering the confidence interval for emission estimates can help the producer to reduce the transaction costs in emissions trading. Comparisons between fuel switch and fuel mix strategies show that fuel mix can provide good tradeoff between profit-making and emissions reduction.     

Controlled predatory pricing in a multiperiod Stackelberg game: an MPEC approach

We analyze a multiperiod oligopolistic market where each period is a Stackelberg game between a leader firm and multiple follower firms. The leader chooses his production level first, taking into account the reaction of the followers. Then, the follower firms decide their production levels after observing the leader’s decision. The difference between the proposed model and other models discussed in literature is that the leader firm has the power to force the follower firms out of business by preventing them from achieving a target sales level in a given time period. The leader firm has an incentive to lower the market prices possibly lower than the Stackelberg equilibrium in order to push the followers to sell less and eventually go out of business. Intentionally lowering the market prices to force competitors to fail is known as predatory pricing, and is illegal under antitrust laws since it negatively affects consumer welfare. In this work, we show that there exists a predatory pricing strategy where the market price is above the average cost and consumer welfare is preserved. We develop a mixed integer nonlinear problem (MINLP) that models the multiperiod Stackelberg game. The MINLP problem is transformed to a mixed integer linear problem (MILP) by using binary variables and piecewise linearization. A cutting plane algorithm is used to solve the resulting MILP. The results show that firms can engage in predatory pricing even if the average market price is forced to remain higher than the average cost. Furthermore, we show that in order to protect the consumers, antitrust laws can control predatory pricing by setting rules on consumer welfare.     

Optimal Pricing and Production Planning for Multi-product Multi-period Systems with Backorders

In this paper, we develop models for production planning with coordinated dynamic pricing. The application that motivated this research is manufacturing pricing, where the products are non-perishable assets and can be stored to fulfill the future demands. We assume that the firm does not change the price list very frequently. However, the developed model and its solution strategy have the capability to handle the general case of manufacturing systems with frequent time-varying price lists. We consider a multi-product capacitated setting and introduce a demand-based model, where the demand is a function of the price. The key parts of the model are that the planning horizon is discrete-time multi-period, and backorders are allowed. As a result of this, the problem becomes a nonlinear programming problem with the nonlinearities in both the objective function and some constraints. We develop an algorithm which computes the optimal production and pricing policy on a finite time horizon. We illustrate the application of the algorithm through a detailed numerical example.     

Short-term manpower planning for MRT carriage maintenance under mixed deterministic and stochastic demands

The purpose of this research is to develop two manpower supply planning models and a solution algorithm for mass rapid transit carriage maintenance under mixed deterministic and stochastic demands. These models are formulated as mixed integer programs that are characterized as NP-hard. We employ problem decomposition techniques, coupled with the CPLEX mathematical programming solver, to develop an algorithm that is capable of efficiently solving the problems. The models and the method used currently in actual operations are evaluated by a simulation-based evaluation method. Finally, we perform a case study using real operating data from a Taiwan MRT maintenance facility. The preliminary results are good, showing that the models could be useful for planning carriage maintenance manpower supply.     

A scenario decomposition approach for stochastic production planning in sawmills

This study considers a real world stochastic multi-period, multi-product production planning problem. Motivated by the challenges encountered in sawmill production planning, the proposed model takes into account two important aspects: (i) randomness in yield and in demand; and (ii) set-up constraints. Rather than considering a single source of randomness, or ignoring set-up constraints as is typically the case in the literature, we retain all these characteristics while addressing real life-size instances of the problem. Uncertainties are modelled by a scenario tree in a multi-stage environment. In the case study, the resulting large-scale multi-stage stochastic mixed-integer model cannot be solved by using the mixed-integer solver of a commercial optimization package, such as CPLEX. Moreover, as the production planning model under discussion is a mixed-integer programming model lacking any special structure, the development of decomposition and cutting plane algorithms to obtain good solutions in a reasonable time-frame is not straightforward. We develop a scenario decomposition approach based on the progressive hedging algorithm, which iteratively solves the scenarios separately. CPLEX is then used for solving the sub-problems generated for each scenario. The proposed approach attempts to gradually steer the solutions of the sub-problems towards an implementable solution by adding some penalty terms in the objective function used when solving each scenario. Computational experiments for a real-world large-scale sawmill production planning model show the effectiveness of the proposed solution approach in finding good approximate solutions.     

带有模糊参数的农业生产计划模型

在现实的生产系统中, 由于材料价格, 产品价格, 市场需求以及劳动者能力等不确定因素的影响, 生产计划问题常常是一个不确定规划问题. 因此, 带有常系数的生产计划模型不能准确有效的描述生产决策环境. 基于可信性理论, 本文将提出一类新的带有模糊参数的生产计划模型. 然后, 我们讨论了可信性函数的逼近并且设计一个基于逼近方法、神经网络和遗传算法的启发式算法来求解这个模糊生产计划问题. 最后, 给出了一个数值例子来表明所设计算法的可行性和有效性.     

Market power analysis in electricity markets using supply function equilibrium model

This paper presents a surveillance method based on the gametheory which is used by the ISO to find whether a power supplierin an electricity market has market power. The paper uses thesupply function equilibrium model to analyse the generationsuppliers’ bidding behaviour and models the ISO's marketpower monitoring problem as a bi-level multi-objective problem.The outer sub-problem is a multi-objective problem which maximizessuppliers’ payoffs, while the inner one is the ISO's marketclearing problem based on the locational marginal pricing mechanism.A discrete method is adopted to find ‘good enough’solutions, in a continuous bidding strategy space, which arethe intersection of all suppliers’ optimal response spacesaccording to Nash equilibrium. The paper utilizes the IEEE 118-bussystem to illustrate the application of the proposed methodwith three suppliers as price setters in the energy market andthe other generators as price takers. The numerical resultsshow that the transmission congestion may enhance the suppliers’ability to exercise market power. Likewise, suppliers’gaming behaviour could relieve the transmission congestion.It is shown that applying price caps is an efficient way ofmitigating market power.     

The deterministic multi-item dynamic lot size problem with joint business volume discount

This paper considers the multi-item dynamic lot size model where joint business volume discount is applied for all items purchased whenever the total dollar value of an order reaches a certain level. Multi-item discounts are prevalent in practical applications, yet the literature has only considered limited instances of single-item models. We establish the mathematical formulation and design an effective dynamic programming based heuristic. Computational results disclose our approach obtains high quality solutions that dominate the best known heuristic for the simplified one-item case, and that proves vastly superior to the state-of-the-art CPLEX MIP code for the multi-item case (for which no alternative heuristics have been devised). We obtained significantly better solutions than CPLEX for the more complex problems, while running from 4800 to over 100,000 times faster. Enhanced variants of our method improve these outcomes further. This revised version was published online in June 2006 with corrections to the Cover Date.     

Dispatch planning using newsvendor dual problems and occupation times: Application to hydropower

We consider price-driven dispatch planning under price uncertainty: A storable commodity is optimally sold and purchased over time. First, we consider models where the storage level is constrained in expectation. The dual of the corresponding optimization problem is related to the newsvendor problem. Exact solutions of bang-bang type are given. The second methodology is for high-frequency dispatch decisions in multistage stochastic programming models: To overcome the curse of dimensionality, prices are modeled by occupation times at price levels. In a case study, we consider a pumped-storage hydropower plant: Numerical solutions are given, which have similar patterns as for the first, exactly solvable problems.     

The transportation problem with exclusionary side constraints and two branch-and-bound algorithms

The transportation problem with exclusionary side constraints, a practical distribution and logistics problem, is formulated as a 0–1 mixed integer programming model. Two branch-and-bound (B&B) algorithms are developed and implemented in this study to solve this problem. Both algorithms use the Driebeek penalties to strengthen the lower bounds so as to fathom some of the subproblems, to peg variables, and to guide the selection of separation variables. One algorithm also strongly exploits the problem structure in selecting separation variables in order to find feasible solutions sooner. To take advantage of the underlying network structure of the problem, the algorithms employ the primal network simplex method to solve network relaxations of the problem. A computational experiment was conducted to test the performance of the algorithms and to characterize the problem difficulty. The commercial mixed integer programming software CPLEX and an existing special purpose algorithm specifically designed for this problem were used as benchmarks to measure the performance of the algorithms. Computational results show that the new algorithms completely dominate the existing special purpose algorithm and run from two to three orders of magnitude faster than CPLEX.     

Envelope Functions: Unifications and Further Properties

Forward–backward and Douglas–Rachford splitting are methods for structured nonsmooth optimization. With the aim to use smooth optimization techniques for nonsmooth problems, the forward–backward and Douglas–Rachford envelopes where recently proposed. Under specific problem assumptions, these envelope functions have favorable smoothness and convexity properties and their stationary points coincide with the fixed-points of the underlying algorithm operators. This allows for solving such nonsmooth optimization problems by minimizing the corresponding smooth convex envelope function. In this paper, we present a general envelope function that unifies and generalizes existing ones. We provide properties of the general envelope function that sharpen corresponding known results for the special cases. We also present a new interpretation of the underlying methods as being majorization–minimization algorithms applied to their respective envelope functions.