Exploring Deep Reinforcement Learning for Task Dispatching in Autonomous On-Demand Services

2021 ◽  
Vol 15 (3) ◽  
pp. 1-23
Author(s):  
Lei Yang ◽  
Xi Yu ◽  
Jiannong Cao ◽  
Xuxun Liu ◽  
Pan Zhou

Autonomous on-demand services, such as GOGOX (formerly GoGoVan) in Hong Kong, provide a platform for users to request services and for suppliers to meet such demands. In such a platform, the suppliers have autonomy to accept or reject the demands to be dispatched to him/her, so it is challenging to make an online matching between demands and suppliers. Existing methods use round-based approaches to dispatch demands. In these works, the dispatching decision is based on the predicted response patterns of suppliers to demands in the current round, but they all fail to consider the impact of future demands and suppliers on the current dispatching decision. This could lead to taking a suboptimal dispatching decision from the future perspective. To solve this problem, we propose a novel demand dispatching model using deep reinforcement learning. In this model, we make each demand as an agent. The action of each agent, i.e., the dispatching decision of each demand, is determined by a centralized algorithm in a coordinated way. The model works in the following two steps. (1) It learns the demand’s expected value in each spatiotemporal state using historical transition data. (2) Based on the learned values, it conducts a Many-To-Many dispatching using a combinatorial optimization algorithm by considering both immediate rewards and expected values of demands in the next round. In order to get a higher total reward, the demands with a high expected value (short response time) in the future may be delayed to the next round. On the contrary, the demands with a low expected value (long response time) in the future would be dispatched immediately. Through extensive experiments using real-world datasets, we show that the proposed model outperforms the existing models in terms of Cancellation Rate and Average Response Time.

Author(s):  
Joelle H. Fong ◽  
Jackie Li

Abstract This paper examines the impact of uncertainties in the future trends of mortality on annuity values in Singapore's compulsory purchase market. We document persistent population mortality improvement trends over the past few decades, which underscores the importance of longevity risk in this market. Using the money's worth framework, we find that the life annuities delivered expected payouts valued at 1.019–1.185 (0.973–1.170) per dollar of annuity premium for males (females). Even in a low mortality improvement scenario, the annuities provide an expected value exceeding 0.950. This suggests that participants in the national annuity pool have access to attractively priced annuities, regardless of sex, product, and premium invested.


2020 ◽  
pp. 1-4
Author(s):  
Andrea Ascoli Marchetti ◽  
Andrea Ascoli Marchetti ◽  
Ciattaglia Riccardo ◽  
Dauri Mario ◽  
Giuliano Ilaria ◽  
...  

Increasing population rates of coronavirus disease 2019 (COVID-19) are occurring in several countries and continents. The impact regarding vascular activity, consequences and complication is scarce and makes the future perspective unclear. The aim of this report is to describe the changes in a high-volume University Hospital, relevant for future decisions. The mortality and morbidity should be higher not only in COVID+ patients but also in vascular patients.


2020 ◽  
Author(s):  
William M. Hayes ◽  
Douglas Wedell

Previous research on experience-based decisions with full feedback supports the idea that people tend to prefer options that minimize the probability of regret. The current study explored whether this preference is modulated by differences in expected value (EV) and the presence or absence of occasional losses. Participants (n = 52) completed an online experiment that involved repeated choices between a safer and a riskier option while receiving full feedback. The riskier option yielded a better outcome on 80% of draws so that choosing it minimized the probability of regret. Preference for the riskier, regret-minimizing option was high when it had the same EV as the safer option and all outcomes were gains, but it decreased when the safer option had a higher EV and when both options included occasional losses. Outcome ratings that were obtained on 50% of trials showed large effects of regret and rejoicing, confirming that participants were sensitive to relative comparisons between obtained and forgone outcomes. Reinforcement-learning modeling indicated that the effects of unequal EVs and mixed outcomes could be accounted for by assuming combined encoding of absolute and relative outcomes and unequal weighting of gains and losses. Overall, these results demonstrate that the impact of regret can be modulated by structural features of the choice environment.


2018 ◽  
Vol 122 (1258) ◽  
pp. 1985-2009 ◽  
Author(s):  
A. Srivastava ◽  
T. St. Clair ◽  
G. Pan

ABSTRACTThe Federal Aviation Administration often blocks strategically located airspace volumes to ensure safety during a variety of operations that are potentially hazardous to aircraft, such as space launches. As the frequency of these operations increases, there is a growing need to deepen collaboration and transparency between stakeholders regarding the use of airspace. This collaboration can be supported by models and capabilities to quickly assess the impact of airspace closures, up to 12 months into the future. This paper presents a technique to enable a ‘what-if’ analysis capability coupled with a prediction model, whereby changes in airspace dimension, location, and activation time are reflected instantaneously as measures of projected impact. The technique can also be used for quick post-operations analysis using historical traffic data and to develop air traffic impact assessment capabilities accessible to a broad range of users outside of the air traffic domain. This research has three key components: developing a model to predict air traffic demand up to 12 months into the future, modelling air traffic impact to the affected traffic, and reducing this information into a data structure that can support on-demand analysis. The focus of this paper is on new techniques to predict demand using a large set of historical track data and further encode these projections to support the quick assessment of the impact of blocking various airspace volumes. Initial results show that the proposed data reduction scheme accurately represented the traffic crossing an airspace and resulted in data size reduction by over 50%. The projection model performed well, the actual number of impacted flights were within the estimated range of approximately 80% of the time. Finally, the responsiveness of the web-based prototype developed to illustrate the concept demonstrated the model’s ability to support an on-demand assessment of the air traffic impact of blocking airspace. A significant limitation of the projection model is that it is based on the historical traffic pattern within the U.S. airspace; separate analysis is needed to adapt it to other geographical location.


2021 ◽  
Vol 285 ◽  
pp. 07019
Author(s):  
Maksim Litvinov

In this paper, we consider the study of the process of seed movement in the seed duct and the impact on the operation of a multi-level seeding system. Mathematical models of the movement of the seed material from the seeder to the plowshare group along the seed line are obtained. To a greater extent, the movement time is affected by the weight of the sown portion and the initial speed of the seeds when falling from the seeder. Also, in straight sections, the time depends on the geometric parameters of the seed line. In the future, the formula of the complete transfer of seeds from the seeder to the plowshare group will allow you to adjust the mode of adjusting the response time of the dosing system on a self-propelled seeder with a multi-level seeding system.


Author(s):  
Priyastiwi Priyastiwi

The purpose of this article is to provide the basic model of Hofstede and Grays’ cultural values that relates the Hofstede’s cultural dimensions and Gray‘s accounting value. This article reviews some studies that prove the model and develop the research in the future. There are some evidences that link the Hofstede’s cultural values studies with the auditor’s judgment and decisions by developing a framework that categorizes the auditor’s judgments and decisions are most likely influenced by cross-cultural differences. The categories include risk assessment, risk decisions and ethical judgments. Understanding the impact of cultural factors on the practice of accounting and financial disclosure is important to achieve the harmonization of international accounting. Deep understanding about how the local values may affect the accounting practices and their impacts on the financial disclosure are important to ensure the international comparability of financial reporting. Gray’s framework (1988) expects how the culture may affect accounting practices at the national level. One area of the future studies will examine the impact of cultural dimensions to the values of accounting, auditing and decision making. Key word : Motivation, leadership style, job satisfaction, performance


Sign in / Sign up

Export Citation Format

Share Document