Mark Sh. Levin

Course “System Design: Structural Approach”

The article describes a course on system design (structural approach) which involves the following: issues of systems engineering; structural models; basic technological problems (structural system modeling, modular design, evaluation/comparison, revelation of bottlenecks, improvement/upgrade, multistage design, modeling of system evolution); solving methods (optimization, combinatorial optimization, multicriteria decision making); design frameworks; and applications. The course contains lectures and a set of special laboratory works. The laboratory works consist in designing and implementing a set of programs to solve multicriteria problems (ranking/selection, multiple choice problem, clustering, assignment). The programs above are used to solve some standard problems (e.g., hierarchical design of a student plan, design of a marketing strategy). Concurrently, each student can examine a unique applied problem from his/her applied domain(s) (e.g., telemetric system, GSM network, integrated security system, testing of microprocessor systems, wireless sensor, corporative communication network, network topology). Mainly, the course is targeted to developing the student skills in modular analysis and design of various multidisciplinary composite systems (e.g., software, electronic devices, information, computers, communications). The course was implemented in Moscow Institute of Physics and Technology (State University).

Towards communication network development (structural systems issues, combinatorial models)

This paper focuses on systems development issues (i.e., improvement, extension) for composite (modular) systems. First, systems approaches are considered: (a) improvement of systems components and/or their interconnections, (b) improvement/extension of a system structure, (c) “space” (e.g., geographical) extension of a system as designing an additional system part, and (d) combined system improvement/extension. Communication networks are examined as a basic applied domain. Second, combinatorial optimization models are pointed out as underlying problems: (i) multicrteria ranking, (ii) knapsack-like problems (e.g., multicrtieria multiple choice problem), (iii) clustering, (iv) multicriteria assignemnt/allocation, etc. Numerical examples for communication networks illustrate the suggested approaches.

Combinatorial evolution of ZigBee protocol

The article addresses combinatorial evolution of communication protocol for wireless sensor networks. Morphological tree structure (a version of and-or tree) is used as a hierarchical model for protocols. Three generations of protocol (ZigBee) are examined. A set of protocol change operations are generated and described. The change operations are used as items for forecasting based on combinatorial problems (e.g., clustering, knapsack problem, multiple choice knapsack problem). Two kinds of forecasts for the examined communication protocol are considered: (i) direct expert based forecast of the protocol, (ii) computed forecast that was based on expert judgment (i.e., assessment of change operations) and usage of multicriteria decision making and combinatorial optimization problems.

Multiset Estimates and Combinatorial Synthesis

This chapter (Partially based on: Levin MS (2012) Multiset estimates and combinatorial synthesis. Electronic preprint, p 30, 9 May 2012, http://arxiv.org/abs/1205.2046 [cs.SY]) addresses new interval multiset estimates and their usage in combinatorial synthesis methods (hierarchical morphological design, multiple choice problem, knapsack problem). The definition of interval multiset estimates, operations over the estimates (integration, comparison, aggregation, alignment), and usage in combinatorial synthesis are described including many numerical examples for various types of the estimates, for the operations over the estimates, for synthesis of three-component and four components systems, for hierarchical synthesis of three-layer hierarchical system.

Multicriteria assignment problem (selection of access points)

This paper addresses assignment of users to access points of wireless network. The considered problem is based on multicriteria assignment model. A set of examined criteria involves the following: (i) maximum of bandwidth, (ii) number of users which are under service at the same time, (iii) network reliability requirements, etc. Two kinds of resource constraints are examined: (a) the number of users under service for each access point, (b) frequency bandwidth that is provided by each access point. The considered optimization problem is NP-hard and heuristic is proposed. Numerical examples illustrate the approach.

Improvement of building automation system

The paper addresses redesign/improvement of a building automation system (BAS). For the sake of simplicity, the field bus technology on KNX example and WSN technologies on IEEE.15.4/Zig.Bee basis are examined. The basic system example consists of four parts: (1) IP/KNX Gateway, (2) IP/WSN 6LoWPAN Gateway, (3) ZigBee Wireless Sensor Network, and (4) KNX Field Bus Infrastructure. A treelike system model (and/or morphological tree) is used. The following system improvement design schemes are examined: (i) upgrade of system components (strategy 1), (ii) extension by adding an additional part (strategy 2), and (iii) combined scheme (strategy 3). Three underlaying problems are used: (a) multicriteria ranking, (b) multicriteria multiple choice problem, and (c) combinatorial synthesis. Numerical examples illustrate the redesign processes.

Towards morphological system design

The article addresses a morphological system design approaches. The following methods are briefly described: (i) basic version of MA, (ii) modification of MA as method of proximity to ideal point(s), (iii) reducing of MA to linear programming, (iv) multicriteria multiple choice problem, (v) quadratic assignment problem, (vi) Pareto-based MA (i.e., revelation of Paretoeffective solutions), (vii) Hierarchical Morphological Multicriteria Design (HMMD) approach, and (viii) Hierarchical Morphological Multicriteria Design (HMMD) approach based on probabilistic/fuzzy estimates. The above-mentioned methods are illustrated by schemes, models, and illustrative examples.

Towards combinatorial evolution of composite systems

The paper describes combinatorial evolution of composite systems. The approach consists of the following stages: (i) design of general hierarchical structure of the examined system; (ii) design of structures for several series system generations; (iii) extraction of changes between the system generations as a set of change items and their evaluation (multicriteria analysis including usage of interval multiset estimates, binary relations over the items: compatibility, complementarity, precedence); (iv) combinatorial synthesis of system forecast(s) as selection of prospective change items while taking into account total constraint(s) (multiple choice problem) or morphological design while taking into account compatibility between the selected change items; (v) aggregation of the obtained system forecast into a resultant forecast. The approach was used for modeling of combinatorial evolution in several applications: (1) DSS COMBI for multicriteria ranking, (2) electronic equipment for image processing, (3) standard for transmission of multimedia information (MPEG, MPEG 2, MPEG 4), and (4) ZigBee communication protocol for wireless sensor network. In the article, the suggested approach is illustrated by evolution of three author courses on multicriteria decision making and modular system design. The following is presented: (a) hierarchical structures of the courses, (b) change items between the courses, (c) combinatorial design of prospective forecasts, (d) aggregation of the forecast into a resultant one.

Towards modular redesign of networked system

The paper describes multicriteria modular approach to redesign/improvement of a networked system (e.g., sensor/ actuator network, communication network). Generally, the approach consists of the following stages: (1) revelation of system bottlenecks, (2) generation of alternative improvement actions for each system bottleneck (i.e., new communication facilities as design alternatives for the upgrade process), (3) multicriteria assessment of the improvement actions, (4) multicriteria ranking of the improvement actions, and (5) forming a generalized improvement plan while taking into account a total resource constraint (multiple choice knapsack problem). A set of requirements/criteria is divided into several groups (criteria clusters: reliability, performance, manageability, etc.). An outranking technique (version of Electre method) is used for multicriteria ranking. A multistage heuristic is used for solving multiple choice knapsack problem. A numerical illustrative example of a real-world cooperative communication network is used to describe the suggested redesign approach. Three redesign schemes are analyzed: (1) a new full system design, (2) a partial system redesign, and (3) a hybrid redesign framework.

Configuration of integrated security system

This paper focuses on modular hierarchical design of configuration for an integrated security system. Hierarchical morphological design approach is used to search for Pareto-effective design solutions. An applied example involves the following: hierarchical (tree-like) model of the security system, design alternatives for system components/parts, criteria for the selection of the design alternatives, solving process as selection of the design alternatives and their synthesis into composite decisions as Pareto-effective solutions. System improvement/reconfiguration and multistage design are considered as well.