Papers by Jiangchuan Huang
2015 European Control Conference (ECC), 2015
We propose the PART-n-TSP policy for the Dynamic Traveling Repairman Problem [1]. We compute a go... more We propose the PART-n-TSP policy for the Dynamic Traveling Repairman Problem [1]. We compute a good approximation for the distribution of the system time, defined as the elapsed time between the arrival and the completion of each task. PART-n-TSP stabilizes the system for every load in [0, 1). PART-n-TSP has lower system time variance than PART-TSP [14] and Nearest Neighbor [1] when the load is neither too small or too large. We show that PART-n-TSP is also optimal for system time expectation under light and heavy loads.
A good model for the dynamic vehicle routing problem is the Dynamic Traveling Repairman Problem (... more A good model for the dynamic vehicle routing problem is the Dynamic Traveling Repairman Problem (DTRP) . The DTRP literature has focused on optimizing the expected value of system time, defined as the elapsed time between the arrival and the completion of each task. We focus on the stability and distribution of system time, including its variance. This paper establishes a partially policy independent necessary and sufficient condition for stability in the DTRP. The policy class includes some of the policies proven to be optimal for system time expectation under light and heavy loads in the literature. We propose a new policy named PART-n-TSP and compute a good approximation for its system time distribution. PART-n-TSP has lower system time variance than PART-TSP and Nearest Neighbor [4] when the load is neither too small or too large. We prove that PART-n-TSP is also optimal for system time expectation under light and heavy loads.
IEEE Conference on Decision and Control and European Control Conference, 2011
We first present a formulation for the real-time tracking of a scalar continuous-time linear proc... more We first present a formulation for the real-time tracking of a scalar continuous-time linear process over an Additive White Gaussian Noise (AWGN) channel without channel feedback and prove several results for minimum mean-squared error (MMSE) tracking. For the one-to-one channel case, we give an optimal encoder-decoder pair along with the optimal tracking performance. With the Gaussian distributed source innovation, one optimal form of the encoder is shown to be a linear innovation encoder, which scales the source innovation to match with the power constraint of the channel input. We then extend the formulation to the case where multiple source processes are tracked via a shared AWGN channel.
We introduce an m-Vehicle Spatial Queuing Problem (m-SQP) extending the m-vehicle Dynamic Traveli... more We introduce an m-Vehicle Spatial Queuing Problem (m-SQP) extending the m-vehicle Dynamic Traveling Repairman Problem (m-DTRP) introduced by Bertsimas and van Ryzin [7]. The m-DTRP is focused on the mean response time (E). We extend it to also consider the variance (V) as a measure of risk, and the distribution of response time where possible. This enables an E −V analysis of system performance to suit individual agencies with different E − V or rewardrisk preferences. We decompose the policy design problem into allocation and sequencing following the literature, but permit a spatial equivalent of only the class of polling policies for allocation. One can restrict attention to this class without loss of optimality in the heavy and light load cases because it includes some of the policies proven to be optimal in the m-DTRP literature. To enable E-V analysis we start by proving that the entire class of spatial polling policies is Markovian, irreducible, and aperiodic at each station. We then formulate a property on the sequencing phase called Economy of Scale (EoS) and show it holds for several policies in the literature. We then prove that the Bertsimas stability condition for the m-DTRP is sufficient for the ergodicity and stability of any sequencing policy satisfying EoS as long as the polling discipline is unlimited. We then describe computational approximations to compute the stationary distributions when the system is stable for some common policies thereby enabling E − V analysis.
We apply virtual machine abstractions to networked autonomous vehicles enabling what we call clou... more We apply virtual machine abstractions to networked autonomous vehicles enabling what we call cloud computing in space. In analogy to traditional system virtualization and cloud computing, there are (customer-operated) virtual vehicles that essentially perform like real vehicles although they are hosted by possibly fewer, shared (provider-operated) real vehicles. Here the focus is, however, on motion rather than computation. In the service-level agreement, a virtual vehicle is a virtual machine plus a virtual speed. We define virtual deadline for each task based on virtual speed, and make the spatial cloud a soft real-time system [7], [1]. The performance isolation is measured by the average of tardiness and delivery probability. We use Voronoi tessellation to allocate the tasks to the real vehicles, and design scheduling policies such as Earliest Virtual Deadline First (EVDF), Earliest Dynamic Virtual Deadline First (EDVDF) and credit scheduling policy for the real vehicles. EVDF is shown to minimize the tardiness. Under EVDF, we identify a worst case arrival process maximizing tardiness. We show in simulation that abstracting real vehicles such as cars or planes to virtual vehicles enables virtual vehicles to move in space like real vehicles with guaranteed tardiness (e.g. ≤ 1%) while being hosted by significantly fewer, e.g. 1-for-7.5, shared real vehicles.
INFORMS Journal on Computing, 2015
We apply virtual machine abstractions to networked autonomous vehicles enabling what we call clou... more We apply virtual machine abstractions to networked autonomous vehicles enabling what we call cloud computing in space. In analogy to traditional system virtualization and cloud computing, there are (customer-operated) virtual vehicles that essentially perform like real vehicles although they are hosted by possibly fewer, shared (provider-operated) real vehicles. Here the focus is, however, on motion rather than computation. In the service-level agreement, a virtual vehicle is a virtual machine plus a virtual speed. We define virtual deadline for each task based on virtual speed, and make the spatial cloud a soft real-time system [7], [1]. The performance isolation is measured by the average of tardiness and delivery probability. We use Voronoi tessellation to allocate the tasks to the real vehicles, and design scheduling policies such as Earliest Virtual Deadline First (EVDF), Earliest Dynamic Virtual Deadline First (EDVDF) and credit scheduling policy for the real vehicles. EVDF is shown to minimize the tardiness. Under EVDF, we identify a worst case arrival process maximizing tardiness. We show in simulation that abstracting real vehicles such as cars or planes to virtual vehicles enables virtual vehicles to move in space like real vehicles with guaranteed tardiness (e.g. ≤ 1%) while being hosted by significantly fewer, e.g. 1-for-7.5, shared real vehicles.
2013 European Control Conference (ECC), 2013
We establish a necessary and sufficient condition for stability in the dynamic traveling repairma... more We establish a necessary and sufficient condition for stability in the dynamic traveling repairman problem (DTRP) [3] under the class of polling-sequencing (P-S) policies satisfying unlimited-polling and economy of scale. The P-S class includes some of the policies proven to be optimal for the expectation of system time under light and heavy loads in the DTRP literature. The number of tasks inside each polling partition is shown to be a Markov chain. Policies such as first come first serve, traveling salesman policy, nearest neighbor and Daganzo's algorithm are shown to have economy of scale.
Trajectory Design via Convex Optimization for Six-Degree-of-Freedom Asteroid Powered Landing
Journal of Guidance, Control, and Dynamics, 2021
This paper presents a convex programming algorithm based on modified Rodrigues parameters for a s... more This paper presents a convex programming algorithm based on modified Rodrigues parameters for a six-degree-of-freedom asteroid powered landing trajectory design problem. The trajectory design probl...
The security of the transportation system depends on that of any of its components and how they a... more The security of the transportation system depends on that of any of its components and how they are interlinked. But the securing of each component is oftentimes in the hand of the agency in whose jurisdiction it falls. Literature on reliability and security economics suggests that when security is defined by the weakest link in an interlinked system, then its level is determined by the agent with the highest cost-benefit ratio, and the other agents have the tendency to under-invest or free ride. When security is a function of total effort, then the opposite obtains and the reliability will depend on the agent with the lowest cost-benefit ratio. These conditions arise in urban transportation. This research explores agency investment behavior in multi-agency urban transportation systems develops guidelines for investments in security. The question to answer: is it preferable to let each agency operate its own security budget and make its own investment decisions or is this process better centralized?
Integrating seismic velocity and well logs with quantified uncertainty
SEG Technical Program Expanded Abstracts 2020, 2020
Optics Express, 2009
Laboratory bidirectional reflectance and polarization measurements were carried out on packed lay... more Laboratory bidirectional reflectance and polarization measurements were carried out on packed layers of both natural sediments and manufactured spherical particles. The results indicate that among the natural sediments showing a strong backscattering peak ("hotspot"), the rough platelets are the only sediments with a branch of negative polarization. Measurements of circular and linear polarization ratios indicate that both smooth ooids and rough platelets are strongly depolarizing. Measurements of perfect spherical grains show both negative polarization and strong backscattering as a remnant of the single scattering process.
Service Provisioning in Cyber-Physical Cloud Computing
We extend the concepts from cloud computing developed in the cyber-world into the physical world,... more We extend the concepts from cloud computing developed in the cyber-world into the physical world, creating a new paradigm called cyber-physical computing cloud. Under the proposed framework, a cyber-physical server is allowed to move in space and perform realworld interactions such as sensing and actuation. The analog for cloud computing's virtual machine, under cyber-physical cloud computing, is the "virtual vehicle". Thus, cyberphysical servers must be hosted by real vehicles, such as automobiles, planes, smartphones, and unmanned air systems. A likely candidate for the application of cyber-physical cloud computing is in air quality monitoring; a collaborative network of vehicle-mounted gas sensors would make possible unique types of atmospheric sensing in three dimensions plus time. We discuss the challenges involved in establishing systems according to this new paradigm, and provide our envisioned solutions, and go on to describe specific applications in air quality and our prototype implementation.
Automate seismic velocity model building through machine learning
SEG Technical Program Expanded Abstracts 2020, 2020
We take the paradigm of cloud computing developed in the cyber-world and put it into the physical... more We take the paradigm of cloud computing developed in the cyber-world and put it into the physical world to create a cyber-physical computing cloud. A server in this cloud moves in space making it a vehicle with physical constraints. Such vehicles also have sensors and actuators elevating mobile sensor networks from a deployment to a service. Possible hosts include cars, planes, people with smartphones, and emerging robots like unmanned aerial vehicles or drifters. We extend the notion of a virtual machine with a virtual speed and call it a virtual vehicle, which travels through space by being bound to real vehicles and by migrating from one real vehicle to another in a manner called cyber-mobility. We discuss some of the challenges and envisioned solutions, and describe our prototype implementation.
Journal of Transportation Security, 2010
The security of the transportation system depends on that of any of its components and how they a... more The security of the transportation system depends on that of any of its components and how they are interlinked. But the securing of each component is oftentimes in the hand of the agency in whose jurisdiction it falls. Literature on reliability and security economics ...
Uploads
Papers by Jiangchuan Huang