Multi-Objective Reinforcement Learning Based Healthcare Expansion Planning Considering Pandemic Events

Abstract

Hospital capacity expansion planning is critical for a healthcare authority, especially in regions with a growing diverse population. Policymaking to this end often requires satisfying two conflicting objectives, minimizing capacity expansion cost and minimizing the number of denial of service (DOS) for patients seeking hospital admission. The uncertainty in hospital demand, especially considering a pandemic event, makes expansion planning even more challenging. This work presents a multi-objective reinforcement learning (MORL) based solution for healthcare expansion planning to optimize expansion cost and DOS simultaneously for pandemic and non-pandemic scenarios. Importantly, our model provides a simple and intuitive way to set the balance between these two objectives by only determining their priority percentages, making it suitable across policymakers with different capabilities, preferences, and needs. Specifically, we propose a multiobjective adaptation of the popular Advantage Actor-Critic (A2C) algorithm to avoid forced conversion of DOS discomfort cost to a monetary cost. Our case study for the state of Florida illustrates the success of our MORL based approach compared to the existing benchmark policies, including a state-of-the-art deep RL policy that converts DOS to economic cost to optimize a single objective.

Demand-Side and Utility-Side Management Techniques for Increasing EV Charging Load

“…Reinforcement learning (RL) techniques are typically preferred for their data-driven online decision-making capability. Recent advances in neural network-based deep RL algorithms lead to widespread applications, including gaming [16], finance [17], energy systems [18], transportation [19], communications [20], environmental systems [21], and healthcare systems [22]. Our preliminary work [18] showcases the efficacy of a deep RL-based predictive maintenance of distribution transformers.…”

“…For this challenging problem with continuous state space and high-dimensional action space, policy gradient RL methods provide an effective solution approach. Specifically, the Advantage Actor-Critic (A2C) algorithm is known to provide quick convergence in such problems [4], [18], [19], [22], [26], [27]. Using two neural networks (actor and critic) A2C reduces the variance of its predecessor policy gradient algorithm the REINFORCE [26].…”

Multi-Objective Advantage Actor-Critic Algorithm for Hybrid Disassembly Line Balancing with Multi-Skilled Workers

“…A typical multi-objective reinforcement learning method is a scalarization method, which constructs scalar rewards according to the combination of competitive rewards, and then applies the single-objective RL algorithm. Salman et al [20] propose a multi-objective Advantage Actor-Critic (A2C) algorithm that sets the balance between two goals by only determining the priority percentage of minimizing capacity expansion cost and minimizing the number of denial of service, making it suitable for decision makers with different abilities, preferences, and needs to solve medical expansion planning solutions. Mohsen [21] proposes a multi-objective reinforcement learning algorithm for solving continuous-valued state-action space without considering different target preferences.…”

Pengaruh Penerapan Program Penyuluhan Kesehatan, Akses Air Minum Bersih, dan Fasilitas Sanitasi terhadap Peningkatan Kesejahteraan Masyarakat di Jawa Barat

“…Program penyuluhan kesehatan, yang ditandai dengan penyediaan layanan kesehatan berbasis masyarakat dan kegiatan promosi kesehatan, telah terbukti memiliki dampak positif terhadap hasil kesehatan, termasuk penurunan penyakit menular dan kematian (Couch & Clow, 2023;Shuvo et al, 2022). Program-program ini bertujuan untuk menyediakan layanan kesehatan dasar dan memberdayakan masyarakat dengan pengetahuan dan sumber daya untuk tindakan pencegahan (Hogg-Graham et al, 2023).…”