Mixed Cost Function Research Articles

Reinforcement Learning Based Diagnosis and Prediction for COVID-19 by Optimizing a Mixed Cost Function From CT Images.

A novel coronavirus disease (COVID-19) is a pandemic disease has caused 4 million deaths and more than 200 million infections worldwide (as of August 4, 2021). Rapid and accurate diagnosis of COVID-19 infection is critical to controlling the spread of the epidemic. In order to quickly and efficiently detect COVID-19 and reduce the threat of COVID-19 to human survival, we have firstly proposed a detection framework based on reinforcement learning for COVID-19 diagnosis, which constructs a mixed loss function that can integrate the advantages of multiple loss functions. This paper uses the accuracy of the validation set as the reward value, and obtains the initial model for the next epoch by searching the model corresponding to the maximum reward value in each epoch. We also have proposed a prediction framework that integrates multiple detection frameworks using parameter sharing to predict the progression of patients' disease without additional training. This paper also constructed a higher-quality version of the CT image dataset containing 247 cases screened by professional physicians, and obtained more excellent results on this dataset. Meanwhile, we used the other two COVID-19 datasets as external verifications, and still achieved a high accuracy rate without additional training. Finally, the experimental results show that our classification accuracy can reach 98.31%, and the precision, sensitivity, specificity, and AUC (Area Under Curve) are 98.82%, 97.99%, 98.67%, and 0.989, respectively. The accuracy of external verification can reach 93.34% and 91.05%. What's more, the accuracy of our prediction framework is 91.54%. A large number of experiments demonstrate that our proposed method is effective and robust for COVID-19 detection and prediction.

Metabolic cost underlies task-dependent variations in motor unit recruitment.

Mammalian skeletal muscles are comprised of many motor units, each containing a group of muscle fibres that have common contractile properties: these can be broadly categorized as slow and fast twitch muscle fibres. Motor units are typically recruited in an orderly fashion following the 'size principle', in which slower motor units would be recruited for low intensity contraction; a metabolically cheap and fatigue-resistant strategy. However, this recruitment strategy poses a mechanical paradox for fast, low intensity contractions, in which the recruitment of slower fibres, as predicted by the size principle, would be metabolically more costly than the recruitment of faster fibres that are more efficient at higher contraction speeds. Hence, it would be mechanically and metabolically more effective for recruitment strategies to vary in response to contraction speed so that the intrinsic efficiencies and contraction speeds of the recruited muscle fibres are matched to the mechanical demands of the task. In this study, we evaluated the effectiveness of a novel, mixed cost function within a musculoskeletal simulation, which includes the metabolic cost of contraction, to predict the recruitment of different muscle fibre types across a range of loads and speeds. Our results show that a metabolically informed cost function predicts favoured recruitment of slower muscle fibres for slower and isometric tasks versus recruitment that favours faster muscles fibres for higher velocity contractions. This cost function predicts a change in recruitment patterns consistent with experimental observations, and also predicts a less expensive metabolic cost for these muscle contractions regardless of speed of the movement. Hence, our findings support the premise that varying motor recruitment strategies to match the mechanical demands of a movement task results in a mechanically and metabolically sensible way to deploy the different types of motor unit.

An optimal control approach to steam distillation of essential oils from aromatic plants

In this work, an optimal steam flow trajectory for essential oils extraction from aromatic plants is derived, minimizing energy consumption. A phenomenological dynamic model of the oil extraction process is adopted from literature and a multi-objective optimal control problem is formulated, in order to minimize energy consumption and at the same time maximize the yield of extraction. The resulting optimal control problem is highly non-linear and it is solved by numerical methods. Simulation results are presented for three scenarios: (i) Maximum yield, (ii) minimum energy consumption and (iii) trade-off between yield and energy. It is shown that the optimal steam flow rate trajectory is not necessarily constant. Using a mixed cost-function, it is possible to extract almost 100% of oil essential while saving 60% of energy. Finally, a sensitivity analysis shows that the optimal steam trajectory has few variations when parameters of the plant physiology change.

An improved imperialist competitive algorithm based photolithography machines scheduling

: Photolithography machine is one of the most expensive equipment in semiconductor manufacturing system, and as such is often the bottleneck for processing wafers. This paper focuses on photolithography machines scheduling with the objective of total completion time minimisation. In contrast to classic parallel machines scheduling, it is characterised by dynamical arrival wafers, re-entrant process flows, dedicated machine constraints and auxiliary resources constraints. We propose an improved imperialist competitive algorithm (ICA) within the framework of a rolling horizon strategy for the problem. We develop a variable time interval-based rolling horizon strategy to decide the scheduling point. We address the global optimisation in every local scheduling by proposing a mixed cost function. Moreover, an adaptive assimilation operator and a sociopolitical competition operator are used to prevent premature convergence of ICA to local optima. A chaotic sequence-based local search method is presented to accelerate the rate of convergence. Computational experiments are carried out comparing the proposed algorithm with ILOG CPLEX, dispatching rules and meta-heuristic algorithms in the literature. It is observed that the algorithm proposed shows an excellent behaviour on cycle time minimisation while with a good on time delivery rate and machine utilisation rate.

Numerically Efficient Self Tuning Strategies for MPC of Integral Gas Engines

AbstractIntegral gas engines represent a critical element in the US gas pipeline network. The stricter emission requirements now facing also these engines have prompted the question for advanced control strategies. Among them MPC has turned out to be a promising choice. Unfortunately, MPC has two main drawbacks a high computational and a demanding and non intuitive tuning effort. These issue become even more critical if speed and precision of the hardware is limited, as it is often the case in embedded solutions. This paper extends previous work on the control of such systems and concentrates on the latter issue, the selftuning. It proposes a tuning approach based on the well-known equivalence between MPC and a set of linear controllers as well as on the use of a mixed cost function which explicitly includes the numerical condition of the MPC problem. As the paper also shows, this yields to an impressive increase in the efficiency and performance of the online algorithm by reducing the number of QP iterations necessary at each step.

An Optimal Control Problem with a Mixed Cost Function

The main objective of this paper is to find an optimal positional control in feedback form for a linear autonomous system with a mixed cost functional. The method of controllability functions is used. The relationship between this problem and an optimal stabilizability problem is also investigated, and several examples are presented.