Prof. Lawrence Mandow
received his MSc degree in Computer Science in 1993 and his PhD in Computer Science in 1999.
He is currently Professor in Computer Science and Artificial
Intelligence at the Computer Sciences and Languages Department of the University of Málaga (Spain).
El Prof. Lawrence Mandow
es Licenciado en Informática (1993) y Doctor Ingeniero en Informática (1999). Actualmente es Catedrático por
el área de Ciencia de la Computación e Inteligencia Artificial en el Departamento de Lenguajes y Ciencias de
la Computación de la Universidad de Málaga (España).
I have co-authored several contributions on the extension of the multiobjective decision making paradigm to graph search algorithms. These include the description and analysis of pathological behaviour in a previous algorithm (MOA*), the development of a provable optimal multiobjective search algorithm (NAMOA*), and a dimensionality reduction technique that greatly enhances the performance of multiobjective best-first search.
I am currently involved in the development of multiobjective generalizations of the Q-learning algorithm. Multiobjective reinforcement learning is a challenging research field. The development of practical algorithms requires an effective way to deal with nonstationary policies. We currently advocate the use of partial models of the environment to control combinatorial growth.
Designing is a complex human endeavour. It requires recognition of opportunities and reflection on the best criteria to achieve the needs and desires that motivate a design process. Therefore, it involves more than problem solving. Multicriteria decision theory provides adequate tools to identify and analyze tradeoffs between criteria. One of the motivations of our research has been the extension of the multicriteria decision paradigm to AI techniques, and its application in engineering and architectural design problems.
Here are some of my recent research projects.
Principal investigator
Funded by: Consejería de Innovación, Ciencia y Empresa, Junta de Andalucía (Spain) – Proyecto de Investigación de Excelencia P07-TIC-03018
Period: 2008-2012.
This project aimed at extending the range and complexity of problems solvable by exact multi-criteria graph search techniques. This included the development of new algorithms and techniques, and the empirical and formal analysis of their properties.
Funded by: Plan Nacional de I+D+i, Ministerio de Ciencia e Innovación (Spain) – TIN2009-14179
Period: 2010-2012
This project aimed at the computational modeling of real-world architectural problems, and the development of decision support techniques based on multiobjective heuristic search and reinforcement learning.
Co-principal investigator (with Prof. José Luis Pérez de la Cruz)
Funded by: Plan Estatal de Investigación Científica y Técnica y de Innovación, Ministerio de Economía y Competitividad (Spain) – TIN2016-80774-R
Period: 2017-2019
This project aims to develop innovative multi-objective reinforcement learning techniques and test their applicability on architectural design problems concerning sustainable design.
Enrique Machuca
An analysis of some algorithms and heuristics for multiobjective graph search
(PhD defence: July 2012)
Francisco J. Pulido Arrebola
New techniques and algorithms for multiobjective and lexicographic goal-based shortest path problems
(PhD defence: July 2015)
Javier Coego Botana
Algoritmos de búsqueda con retroceso para problemas multicriterio (in Spanish)
(PhD defence: December 2015)
