mhpma.m

%Author: Zsolt T. Kosztyan Ph.D habil., University of Pannonia,
%Faculty of Economics, Department of Quantitative Methods
%----------------
%Implementation of hyprid project management agent (HPMa)
%----------------
%Output:
%PSM: N by M+1 matrix of the calculated project plan. PSM contains the
%logic domain, time and cost demands and (depends on the problem selection 
%it contains) quality and/or resource demands. 
%PSM=[DSM,TD,CD{,QD}{,RD},EST|SST]
% DSM is an N by N upper triangular binary matrix of the calculated logic
% domain
% TD is an N by 1 column vector of task durations 
% CD is an N by 1 column vector of cost demands
%depends on the problem selection
% QD is an N by 1 column vector of quality demands
% RD is an N by nR matrix of resource demands
%depends on the resource allocation is performed or not
% SST is an N by 1 column vector of scheduled start time of tasks if
%resource allocation is applied
% EST is an N by 1 column vector of earliest start time of tasks if
%resource demands are not considered
%---------------- 
%Inputs:
%PDM: N by N+(M-N)*W matrix of the stochastic project plan. 
%PDM=[PEM,TD,CD{,QD}{,RD}], where PEM is an N by N upper triangular matrix 
%of logic domain, 
% TD is an N by w matrix of task durations,
% CD is an N by w matrix of cost demands,
%depends on the problem selection
% QD is an N by w matrix of quality parameters,
%depends on the resource allocation is performed or not
% RD is an N by w*nR matrix of resource demands.
%const: 3+nR+1 elements row vector of [Ct,Cc,Cq,CR,Cs] values, where
% Ct is the time constraint (max constraint)
% Cc is the cost constraint (max constraint)
% Cq is the quality constraint (min constraint)
% CR are the resource constraints (max constraint)
% Cs is the score constraint (min constraint)
%CONST: 3..3+nR+1 (depends on the problem selection) elements row vector of 
%[Ct,Cc,{Cq,CR,}Cs] values, where
% Ct is the time constraint (max constraint, required in all problems)
% Cc is the cost constraint (max constraint, required in all problems)
% Cq is the quality constraint (min constraint, required in TQCTP problems))
% CR are the resource constraints (max constraint, required in 
%resource-allocation problems)
% Cs is the score constraint (min constraint, required in all problems)
%Select: 1-18 scalar of problem selection
% 1: APSP: Agile Project Scheduling Problem
% 2: APSPq: Agile Project Scheduling Problem with quality parameters
% 3: HDTCTP: Hybrid Discrete Time-Cost Trade-off Problem
% 4: HCTCTP: Hybrid Continouos Time-Cost Trade-off Problem
% 5: HDTQCTP: Hybrid Discrete Time-Quality-Cost Trade-off Problem
% 6: HCTQCTP: Hybrid Continouos Time-Quality-Cost Trade-off Problem
% 7: RC-APSP: Resource-Constraint Agile Project Scheduling Problem 
% 8: PO-APSP: Pareto-Optimal resource-allocated Agile Project Scheduling 
%Problem 
% 9: RC-HDTCTP: Resource-Constraint Hybrid Discrete Time-Cost Trade-off 
%Problem
%10: RC-HCTCTP: Resource-Constraint Hybrid Continouos Time-Cost Trade-off 
%Problem
%11: PO-HDTCTP: Pareto-Optimal resource-allocated Hybrid Discrete Time-Cost 
%Trade-off Problem
%12: PO-HCTCTP: Pareto-Optimal resource-allocated Hybrid Continouos 
%Time-Cost Trade-off Problem
%13: RC-APSPq: Resource-Constraint Agile Project Scheduling Problem 
%with quality parameters
%14: PO-APSPq: Pareto-Optimal resource-allocated Agile Project Scheduling 
%with quality parameters
%15: RC-HDTQCTP: Resource-Constraint Hybrid Discrete Time-Quality-Cost 
%Trade-off Problem
%16: PO-HDTQCTP: Pareto-Optimal resource-allocated Hybrid Discrete 
%Time-Quality-Cost Trade-off Problem
%17: RC-HCTQCTP: Resource-Constraint Hybrid Continouos Time-Quality-Cost 
%Trade-off Problem
%18: PO-HCTQCTP: Pareto-Optimal resource-allocated Hybrid Continouos 
%Time-Quality-Cost Trade-off Problem
%typetfcn: Type of target function 
% {0=maxTPQ,} 1=minTPT, 2=minTPC, 3=maxTPS, {4=minUF,} ~ composite
%w: Number of modes
%---------------- 
%Usage:
%PSM=mhpma(PDM,CONS,Select,typefcn,w)
%---------------- 
%Example:
%PDM=[triu(rand(10)*.5+.5),20*rand(10,3),30*rand(10,3),rand(10,3),...
%   5*rand(10,6)];
%const=[percentt(PDM,3,.9),percentc(PDM,3,.9),,percentq(PDM,3,.7),
%   percentqr(PDM,3,1)',percents(PDM,.7)];
%Select=15;
%typefcn=999; %let target function be a composite target function
%w=3;
%tic;PSM=mhpma(PDM,const,Select,typefcn,w);toc
%---------------- 
%Prepositions and Requirements:
%1.)The logic domian of the input PDM matrix must be an upper triangular 
%matrix, where the matrix elements are between 0 to 1 interval.
%2.)At least one matrix element is lower than 1 and upper than 0. => There
%are at least one decision variable
%3.)The number of modes(w) is a positive integer. In case of countinous
%trade-off problems w=2 and in case of selections the w shuld be 1.
%4.)The number of resources (nR) is a positive number
%5.)The elements of Time/Cost/Quality/Resource Domains are positive real 
%numbers.
%6.)Usually a monotonity are assumed, which, means: if tk,i<tk,j => 
%ck,i>=ck,j and qk,i<=qk,j, where 1<=k<=N is the k-th task, 1<=i,j<=w are 
%the selected modes. This asumptation is only used in continous trade-off
%problems. 
%7.)HPMa can change two kinds of variables: a.) flexible task dependencies
%and uncertain task realizations; b.) completion modes. If the logic domain
%is fixed, HPMa=TPMa, while if the number of modes, w=1, HPMa=APMa.

function PSM=mhpma(PDM,CONS,Select,typefcn,w)
N=size(PDM,1); %Number of activities
PEM=PDM(:,1:N); %PEM is the first N by N domain of logic plan
M=size(PDM,2); %M is the number of activities + modes of activities, costs, 
%(quality parameters, resources (depends on the problem))
switch Select %Problem selection
    case 1 % No resources, select normal parameters, no quality parameters;
        %select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost demands) are
        %selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        CD=zeros(N,1);
        MODES=MODES';
        for I=1:N %The cost of normal duration is the normal cost
            if MODES(I)>0
                CD(I)=C(I,MODES(I));
            end
        end
        PDM=[PEM,TD,CD];
        PSM=apmgen(PDM,CONS,typefcn); %In this case HPMa=APMa
   case 2 % No resources, select normal parameters, with quality parameters;
        %select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost demands, quality 
        %parameters) are selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        Q=PDM(:,N+2*w+1:N+3*w);
        CD=zeros(N,1);
        QD=zeros(N,1);
        MODES=MODES';
        for I=1:N
            if MODES(I)>0
                CD(I)=C(I,MODES(I));
                QD(I)=Q(I,MODES(I));
            end
        end % The cost of normal duration is the normal cost. Similarly the
        %the normal quality parameteds is the quality paremeters of normal
        %durations
        PDM=[PEM,TD,CD,QD];
        PSM=apmgenq(PDM,CONS,typefcn); %In this case HPMa=APMa
    case 3 % No resources, dtctp, no quality parameters;
        PSM=hpmgend(PDM,CONS,typefcn);
    case 4 % No resources, ctctp, no quality parameters;
        PSM=hpmgench(PDM,CONS,typefcn);   
    case 5 % No resources, dtctp, with quality parameters;
        PSM=hpmgendq(PDM,CONS,typefcn);
    case 6 % No resources, ctctp, with quality parameters;
        PSM=hpmgencq(PDM,CONS,typefcn); 
    case 7 % With resources, select normal parameters, no quality parameters;
        %select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost demands, quality 
        %parameters) are selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        R=PDM(:,N+2*w+1:end);
        nR=(M-N-2*w)/w;
        CD=zeros(N,1);
        RD=zeros(N,nR);
        MODES=MODES';
        for I=1:N % The cost/resource demands of normal duration is the 
            %normal cost/resources
            if MODES(I)>0
                CD(I)=C(I,MODES(I));
                for J=1:nR
                    RD(I,J)=R(I,(J-1)*MODES(I)+1);
                end
            end
        end
        PDM=[PEM,TD,CD,RD];
        PSM=hpmgendr(PDM,CONS,typefcn); %In this case HPMa=APMa
    case 8 % Pareto-optimal resource allocation, select normal parameters, 
        %no quality parameters; select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost, resource 
        %demands) are selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        R=PDM(:,N+2*w+1:end);
        nR=(M-N-2*w)/w;
        CD=zeros(N,1);
        RD=zeros(N,nR);
        MODES=MODES';
        for I=1:N %The cost/resource demands of normal duration is the
            %normal cost/resources
            if MODES(I)>0
                CD(I)=C(I,MODES(I));
                for J=1:nR
                    RD(I,J)=R(I,(J-1)*MODES(I)+1);
                end
            end
        end
        PDM=[PEM,TD,CD,RD];
        PSM=hpmgendpr(PDM,CONS,typefcn,1); %In this case HPMa=APMa
    case 9 % With resources, dtctp, no quality parameters;
        PSM=hpmgendr(PDM,CONS,typefcn); 
    case 10 % With resources, ctctp, no quality parameters;
        PSM=hpmgencr(PDM,CONS,typefcn);
    case 11 % Pareto-optimal resource allocation, dtctp, no quality 
        %parameters;
        PSM=hpmgendpr(PDM,CONS,typefcn,w); 
    case 12 % Pareto-optimal resource allocation, ctctp, no quality
        %parameters;
        PSM=hpmgenchpr(PDM,CONS,typefcn);
    case 13 % With resources, select normal parameters, with quality 
        %parameters; select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost demands, quality 
        %parameters, resource demands) are selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        Q=PDM(:,N+2*w+1:N+3*w);
        R=PDM(:,N+3*w+1:end);
        nR=(M-N-3*w)/w;
        CD=zeros(N,1);
        QD=zeros(N,1);
        RD=zeros(N,nR);
        MODES=MODES';
        for I=1:N
            if MODES(I)>0 % The cost/quality/resource demands of normal 
            %duration is the normal cost/resources/quality parameters
                CD(I)=C(I,MODES(I));
                QD(I)=Q(I,MODES(I));
                for J=1:nR
                    RD(I,J)=R(I,(J-1)*MODES(I)+1);
                end
            end
        end
        PDM=[PEM,TD,CD,QD,RD];
        PSM=hpmgendqr(PDM,CONS,typefcn); %In this case HPMa=APMa
    case 14 % Pareto-optimal resource allocation, select normal parameters, with quality parameters;
        %select variables
        %Despite w=1 is assumed, this section works, if w>1, however, in
        %this case normal paramteres (i.e. duration, cost demands, quality 
        %parameters) are selected to evaluate
        [TD,MODES]=max(PDM(:,N+1:N+w)'); %Normal=maximal duration is 
        %selected.
        TD=TD';
        C=PDM(:,N+w+1:N+2*w);
        Q=PDM(:,N+2*w+1:N+3*w);
        R=PDM(:,N+3*w+1:end);
        nR=(M-N-3*w)/w;
        CD=zeros(N,1);
        QD=zeros(N,1);
        RD=zeros(N,nR);
        MODES=MODES';
        for I=1:N % The cost/quality/resource demands of normal 
            %duration is the normal cost/resources/quality parameters
            if MODES(I)>0
                CD(I)=C(I,MODES(I));
                QD(I)=Q(I,MODES(I));
                for J=1:nR
                    RD(I,J)=R(I,(J-1)*MODES(I)+1);
                end
            end
        end
        PDM=[PEM,TD,CD,QD,RD];
        PSM=hpmgendqpr(PDM,CONS,typefcn,1); %In this case HPMa=APMa
    case 15 % With resources, dtctp, with quality parameters;
        PSM=hpmgendqr(PDM,CONS,typefcn); 
    case 16 % Pareto-optimal resource allocation, dtctp, with quality parameters;
        PSM=hpmgendqpr(PDM,CONS,typefcn,w);
    case 17 % With resources, ctctp, with quality parameters;
        PSM=hpmgencqr(PDM,CONS,typefcn); 
    case 18 % Pareto-optimal resource allocation, ctctp, with quality parameters;
        PSM=hpmgencqpr(PDM,CONS,typefcn);
    case 19 % Discrete version of pareto-optimal multi-mode resource constrained project 
        %scheduling problem without quality parameters
        PSM=parhpmgendr(PDM,CONS);
    case 20 % Continouos version of pareto-optimal multi-mode resource constrained project 
        %scheduling problem without quality parameters
        PSM=parhpmgencr(PDM,CONS);
    case 21 % Discrete version of pareto-optimal multi-mode resource constrained project 
        %scheduling problem with quality parameters
        PSM=parhpmgendqr(PDM,CONS);
    case 22 % Continouos version of pareto-optimal multi-mode resource constrained project 
        %scheduling problem with quality parameters
        PSM=parhpmgencqr(PDM,CONS);
    otherwise
        PSM=PDM;
end