optimal decision

We describe the use of a successful combination of Bayesian inference and decision theory in a clinical trial design. The trial involves three important decisions, adaptive dose allocation, optimal stopping of the trial, and the optimal terminal decision upon stopping. For all three decisions we use a formal Bayesian decision-theoretic approach. The application demonstrates how Bayesian posterior inference and decision-theoretic approaches combine to provide a coherent solution in a complex application. The main challenges are the need for a flexible probability model for the unknown dose-response curve, a delayed response, the sequential nature of the stopping decision, and the complex considerations involved in the terminal decision. The main methodological features of the proposed solution are the use of decision theory to achieve optimal learning about the unknown dose-response curve, an innovative grid-based approximation method to implement backward induction for the sequentia...

In this article, a general problem of sequential statistical inference for general discrete-time stochastic processes is considered. The problem is to minimize an average sample number given that Bayesian risk due to incorrect decision does not exceed some given bound. We characterize the form of optimal sequential stopping rules in this problem. In particular, we have a characterization of the form of optimal sequential decision procedures when the Bayesian risk includes both the loss due to incorrect decision and the cost of observations.

In this article, a general problem of sequential statistical inference for general discrete-time stochastic processes is considered. The problem is to minimize an average sample number given that Bayesian risk due to incorrect decision does not exceed some given bound. We characterize the form of optimal sequential stopping rules in this problem. In particular, we have a characterization of the form of optimal sequential decision procedures when the Bayesian risk includes both the loss due to incorrect decision and the cost of observations.

This paper evaluates the simultaneous determination of price and inventory replenishment when a firm faces demand from distinct market segments. A firm utilizes fences, such as advance or nonrefundable payment, to maintain separation of its market segments; however, fences are imperfect and allow a degree of demand leakage from the higher-priced to the lower-priced market segment. We investigate the optimal structure of joint price and inventory decisions with fencing, and demonstrate that more segments is not necessarily better, especially when demand uncertainty is high in the presence of lost sales. We also show the impact of imperfect fences on the firm's profitability, and evaluate how fencing costs affect the optimal fencing decision.

We describe a modelling and problem-solving process supported by a decision support system, which accounts for the possibility of different models depending on factors like multiple conflicting objectives, uncertainty about the consequences of the decisions, imprecision concerning the decision maker's preferences and decisions with consequences that span several time periods. This decision support system addresses all the above aspects, which can be simultaneously taken into account in real decision making problems, by considering an objective hierarchy that models appropriate time preferences. The analysis is focused on judgemental inputs, imprecise scaling factors on objectives and imprecise scalar utility functions on the attributes of the hierarchy for the purpose of aggregation, if appropriate, in a scalar or vectorial global additive utility function, which can incorporate discount factors to evaluate the timestream vectors. The decision support system includes an optimisation method based on the simulated annealing metaheuristic, as well as a multiparametric sensitivity analysis as an aid for identifying the optimal alternative.

Following work of Stroud and Saeger, we investigate the formulation of the port of entry inspection algorithm problem as a problem of finding an optimal binary decision tree for an appropriate Boolean decision function. We report on an experimental analysis of the robustness of the conclusions of the StroudSaeger analysis and show that the optimal inspection strategy is re markably insensitive to variations in the parameters needed to apply the StroudSaeger method.

In real-life situations involving risk and uncertainty, optimal policy hinges on selecting a course of action characterized by the highest expected value (i.e., future outcomes weighted by their probabilities). Nevertheless, a vast body of findings from economic and psychological studies indicate that people rarely follow this principle1–3. The attempts to make optimal decisions are impeded by the complexity of a task and the computational capabilities of decision makers, often leading to suboptimal choices4. Recent research has demonstrated that decision makers are systematically biased toward suboptimal options5. However, little is known about the nature of this bias. Here we show that recurring suboptimal choices result in superior decision making. In one simulation study and three well-powered (N = 1,046) fully-incentivized empirical studies employing a task to mimic decision making under uncertainty, we demonstrated that people who traded off their decision accuracy for the num...

With growing interest in recovering materials and subassemblies within consumer products at the end of their useful life, there has been an increasing interest in developing decision-making methodologies that determine how to maximize the environmental benefits of end-oflife (EOL) processing while minimizing costs under variable EOL situations. This paper describes a methodology to analyze how product designs and situational variables impact the Pareto set of optimal EOL strategies with the greatest environmental benefit for a given economic cost or profit. Since the determination of this Pareto set via enumeration of all disassembly sequences and EOL fates is prohibitively time-consuming even for relatively simple products, multi-objective genetic algorithms (GA) are utilized to rapidly approximate the Pareto set of optimal EOL tradeoffs between cost and environmentally conscious actions. Such rapid calculations of the Pareto set are critical to better understand the influence of situational variables on how disassembly and recycling decisions change under different EOL scenarios (e.g., under variable regulatory, infrastructure, or market situations). To illustrate the methodology, a case study involving the EOL treatment of a coffee maker is described. Impacts of situational variables on tradeoffs between recovered energy and cost in Aachen, Germany, and in Ann Arbor, MI, are elucidated, and a means of presenting the results in the form of a multi-situational EOL strategy graph is described. The impact of the European Union Directive regarding Waste Electric and Electronic Equipment (WEEE) on EOL trade-offs between energy recovery and cost was also considered for both locations.

The problem of calculating the local and global decision thresholds in hard decisions based cooperative spectrum sensing is well known for its mathematical intractability. Previous work relied on simple suboptimal counting rules for decision fusion in order to avoid the exhaustive numerical search required for obtaining the optimal thresholds. However, these simple rules are not globally optimal as they do not maximize the overall global detection probability by jointly selecting local and global thresholds. Instead, they maximize the detection probability for a specific global threshold. In this paper, a globally optimal decision fusion rule for Primary User signal detection based on the Neyman-Pearson (NP) criterion is derived. The algorithm is based on a novel representation for the global performance metrics in terms of the regularized incomplete beta function. Based on this mathematical representation, it is shown that the globally optimal NP hard decision fusion test can be put in the form of a conventional one dimensional convex optimization problem. A binary search for the global threshold can be applied yielding a complexity of O(log 2 (N )), where N represents the number of cooperating users. The logarithmic complexity is appreciated because we are concerned with dense networks, and thus N is expected to be large. The proposed optimal scheme outperforms conventional counting rules, such as the OR, AND, and MAJOR-ITY rules. It is shown via simulations that, although the optimal rule tends to the simple OR rule when the number of cooperating secondary users is

We propose a method for computing the range of the optimal decisions when the utility function runs through a class U. The class U has constraints on the values and the shape of the utility functions. A discretization method enables to easily approximate the optimal decision associated with a particular utility function u 2 U. The range of optimal decisions is computed by a Monte Carlo optimization method. An example is provided with numerical results.

Possibilistic de cision theory has be e n propose d twe nty years ago and has had se ve ral extensions since the n. Even though ap pe aling for its ability to handle qualitative decision proble ms, possibilistic decision the ory suffe rs from an important drawback. Qualitative possibilistic utility criteria compare acts through min and max operators, which Jeads to a drowning effect. To over come this Jack of decision power of the theory, several refinements have been proposed. Lexicographie refinements are particularly appealing since they allow to benefit from the Expected Utility background, while remaining qualitative. This article aims at extend ing lexicographie refinements to sequential decision problems i.e., to possibilistic decision trees and possibilistic Markov decision proce sse s, whe n the horizon is finite . We present two crite ria that re fine qualitative possibilistic utilitie s and provide dynarnic prograrnming algorithrns for calculating Jexicographically optimal policies.

We develop a Bayesian model for decision-making under time pressure with endogenous information acquisition. In our model, the decision-maker decides when to observe (costly) information by sampling an underlying continuoustime stochastic process (time series) that conveys information about the potential occurrence/non-occurrence of an adverse event which will terminate the decisionmaking process. In her attempt to predict the occurrence of the adverse event, the decision-maker follows a policy that determines when to acquire information from the time series (continuation), and when to stop acquiring information and make a final prediction (stopping). We show that the optimal policy has a "rendezvous" structure, i.e. a structure in which whenever a new information sample is gathered from the time series, the optimal "date" for acquiring the next sample becomes computable. The optimal interval between two information samples balances a trade-off between the decision maker's "surprise", i.e. the drift in her posterior belief after observing new information, and "suspense", i.e. the probability that the adverse event occurs in the time interval between two information samples. Moreover, we characterize the continuation and stopping regions in the decisionmaker's state-space, and show that they depend not only on the decision-maker's beliefs, but also on the "context", i.e. the current realization of the time series.

Mobile Ad Hoc Networks (MANET) are a peer-topeer communication network that can be used to transmit data using a mobile or wireless link without the support of fixed infrastructure. A literature survey highlighted that the key characteristics of efficient MANET nodes are reputation value, residual battery level and energy consumption. Various MANET algorithms and routing techniques such as location aided routing, broadcasting, multicasting and energy efficient routing have been developed. MANET intelligent routing algorithm requirements include a range of characteristics. This paper proposes a new MANET routing algorithm that includes quadrant based opportunistic routing, an intelligent energy matrix and energy status request messages with packet receipt acknowledgement notification. In quadrant based routing packets are transmitted towards the quadrant that the destination resides within. The proposed algorithm uses an intelligent energy matrix that creates a look up table including the key characteristics: reputation value, residual battery level and energy consumption. The proposed algorithm balances the traffic uniformly across four intermediate nodes in any desired quadrant. The simulation results presented in this paper demonstrate that due to the inclusion of the energy matrix and quadrant based routing, the number of broadcast messages decreases, reducing data flooding, providing improved channel efficiency and improves bandwidth utilization. Load balancing also increases the lifetime of intermediate nodes which provides improved route stability. The proposed algorithm also includes packet receipt acknowledgement notification which guarantees end-to-end packet delivery and provides reliable MANET routing. Comparison with two other MANET broadcast algorithms identifies that a slower transmission speed is offset by the proposed routing algorithm providing higher energy efficiency, reliability and stability.

Decision trees and decision rule systems play important roles as classifiers, knowledge representation tools, and algorithms. They are easily interpretable models for data analysis, making them widely used and studied in computer science. Understanding the relationships between these two models is an important task in this field. There are well-known methods for converting decision trees into systems of decision rules. In this paper, we consider the inverse transformation problem, which is not so simple. Instead of constructing an entire decision tree, our study focuses on a greedy polynomial time algorithm that simulates the operation of a decision tree on a given tuple of attribute values.

Decision theory requires agents to assign probabilities to states of the world and utilities to the possible outcomes of different actions. When agents commit to having the probabilities and/or utilities in a decision problem defined by objective features of the world, they may find themselves unable to decide which actions maximize expected utility. Decision theory has long recognized that work-around strategies are available in special cases; this is where dominance reasoning, minimax, and maximin play a role. Here we describe a different work around, wherein a rational decision about one decision problem can be reached by "interpolating" information from another problem that the agent believes has already been rationally solved.

When the benefit of making a correct decision is sufficiently high, even a slight increase in the probability of making such a decision justifies an increase in the number of decision makers. Applying a standard uncertain dichotomous choice benchmark setting, this study focuses on the relative desirability of two alternatives: adding individuals with capabilities identical to the existing ones and adding identical individuals with mean-preserving capabilities that depend on the states of nature. Our main result establishes that when the group applies the simple majority rule, variability in the capabilities of the new decision makers under the two states of nature, which is commonly observed in various decision-making settings, is less desirable in terms of the probability of making the correct decision.

A decision science blind to decision procedures would be ”unfair”: The effect of decision process on decision-outcome satisfaction and subsequent choice in a performance environment Daniel DeCaro Miami University (MU) Joseph Johnson Mimia Univeristy (MU) Abstract: Contemporary models of decision making assume individuals evaluate options solely in terms of their ex- pected outcomes. However, recent research indicates that in institutional settings decision makers are also concerned with the procedural fairness of the decision process that generates those outcomes, such as whether decision makers were granted democratic inclusion in the decision-making process itself. We provide a much-needed quantification of the value decision makers place on inclusive decision procedures, showing (a) the pattern by which decision proce- dures alter individuals’ perceptions of otherwise identical outcomes, spanning losses and gains of differing quantity and quality (e.g., failure/success) and (b) t...

Many decisions in life are sequential and constrained by a time window. Although mathematically derived optimal solutions exist, it has been reported that humans often deviate from making optimal choices. Here, we used a secretary problem, a classic example of finite sequential decision-making, and investigated the mechanisms underlying individuals’ suboptimal choices. Across three independent experiments, we found that a dynamic programming model comprising subjective value function explains individuals’ deviations from optimality and predicts the choice behaviors under fewer opportunities. We further identified that pupil dilation reflected the levels of decision difficulty and subsequent choices to accept or reject the stimulus at each opportunity. The value sensitivity, a model-based estimate that characterizes each individual’s subjective valuation, correlated with the extent to which individuals’ physiological responses tracked stimuli information. Our results provide model-ba...

We have proposed a novel interactive procedure for performing decision analysis, called Robust Interactive Decision-Analysis (RID), which avoids the difficult problems in measuring utility and state probability information associated with traditional decision tree analysis. Instead, the RID method permits a decision maker (DM) to voluntarily and interactively express strong (viz, sure) binary preference for actions, partial decision functions, and full decision functions, and only imprecise probability and utility function assessments. These inputs serve as operators to prune the event probability space and decision space until an optimal choice strategy is obtained. Conceptually, the operation of the RID method can be regarded as a state-space pruning system and the computer implementation of the RID methodology can be viewed as a decision support system. Having described the RID methodology and its theoretical developments, we continue the illustration of the corresponding state-space pruning system and its properties. We also discuss the system principles which we have derived from psychological theories and end-user interface studies as well RID method. as our implementation experiences with the

This paper provides an operational overview of the principle features and issues for Robust Interactive Decision-Analysis (RID), an alternative to traditional decision tree analysis. The RID method avoids the difficult problems in measuring utility and state probability information associated with decision tree analysis. Instead, this alternative approach elicits only some imprecise (ordinal or interval) probability estimates and strong preference binary choices and incorporates them into an interactive procedure which repeatedly prunes the set of possible decision functions until a single 'optimal' or subset of efficient decision functions remains. The paper provides an informal introduction to the RID methodology and also discusses a decision support system which implements these ideas. We also discuss the system design criteria which have evolved from our own implementation experiences with the RID methodology.

It is substantiated that one of the problematic issues of pedagogical testing is the inconsistency (or incomplete compliance) of the structure and content of test tasks, on the one hand, and the information model of knowledge representation about the academic discipline implemented in the learning process, on the other hand. The approach to resolving this issue is set forth where the semantic knowledge network is considered as an information model for representing knowledge about a subject area. The definition of the quasi-fractal graph is introduced, it is shown that the semantic network of knowledge about the subject area, as well as the information model of the pedagogical test, can be represented as quasi-fractal graphs. It is shown that the quasi-fractal graph is one of the forms for recording the semantic knowledge network. The conditions under which the structure and content of the test tasks will correspond to the information model of knowledge representation about the academic discipline being implemented in the learning process.

In this thesis an attempt is made to identify the capabilities and limitations of the human decision-maker in multi-stage decision tasks and to investigate and evaluate methods of aiding him, particularly with the use of on-line predictive displays. In multi-stage decision tasks, the decisions already made by the decision-maker affect the decision-situations he will subsequently face. This requires that experimental studies of such decision-making must present equivalent interactive decision situations. Although the necessary methodology for such studies is rather complicated, the results thereby obtained are more relevant to many real-life situations. [Continues.]

To make fast and accurate behavioral choices, we need to integrate noisy sensory input, take prior knowledge into account, and adjust our decision criteria. It was shown previously that in two-alternative-forced-choice tasks, optimal decision making can be formalized in the framework of a sequential probability ratio test and is then equivalent to a diffusion model. However, this analogy hides a "chicken and egg" problem: to know how quickly we should integrate the sensory input and set the optimal decision threshold, the reliability of the sensory observations must be known in advance. Most of the time, we cannot know this reliability without first observing the decision outcome. We consider here a Bayesian decision model that simultaneously infers the probability of two different choices and at the same time estimates the reliability of the sensory information on which this choice is based. We show that this can be achieved within a single trial, based on the noisy responses of sensory spiking neurons. The resulting model is a non-linear diffusion to bound where the weight of the sensory inputs and the decision threshold are both dynamically changing over time. In difficult decision trials, early sensory inputs have a stronger impact on the decision, and the threshold collapses such that choices are made faster but with low accuracy. The reverse is true in easy trials: the sensory weight and the threshold increase over time, leading to slower decisions but at much higher accuracy. In contrast to standard diffusion models, adaptive sensory weights construct an accurate representation for the probability of each choice. This information can then be combined appropriately with other unreliable cues, such as priors. We show that this model can account for recent findings in a motion discrimination task, and can be implemented in a neural architecture using fast Hebbian learning.

In C situations, decision makers would love having a crystal ball that describes future events not under their control, and how these events would affect each of the courses of action (COAs) being considered. Further, decision makers would appreciate knowing the potential cost of each COA—e.g, via a metric based on the numbers of resources used and the damage, injury, and deaths that may result. While crystal ball technology remains as elusive as ever, Lempert et al. and Chandrasekaran have developed general methods for identifying robust COAs by using simulation models that determine the plausible consequences of each COA under a wide range of possible futures. Because the simulation models must be run many times for each of the (possibly many) COAs, these techniques are computationally intensive, sometimes taking hours, days or weeks. Since tactical commanders need to make decisions in minutes or seconds, we have been manipulating the models underlying two different simulations to...

We present the design and implementation of a custom discrete optimization technique for building rule lists over a categorical feature space. Our algorithm provides the optimal solution, with a certi cate of optimality. By leveraging algorithmic bounds, e cient data structures, and computational reuse, we achieve several orders of magnitude speedup in time and a massive reduction of memory consumption. We demonstrate that our approach produces optimal rule lists on practical problems in seconds. This framework is a novel alternative to CART and other decision tree methods. CCS CONCEPTS • Computing methodologies → Discrete space search; Classi cation and regression trees;

It is now u n d e r s t o o d [2, 3 1 t h a t the t u r b o decoding algorithm is a n instance of a probability propagation a l g o r i t h m (P P A) o n a g r a p h w i t h m a n y cycles. I n this paper w e investigate t h e behavior of a n P P A in graphs w i t h a single cycle such as the g r a p h of a tail-biting code. First, w e show t h a t for strictly positive local kernels, t h e iterations of the

A greedy algorithm has been presented in this paper to construct decision trees for three different approaches (many-valued decision, most common decision, and generalized decision) in order to handle the inconsistency of multiple decisions in a decision table. In this algorithm, a greedy heuristic 'misclassification error' is used which performs faster, and for some cost function, results are better than 'number of boundary subtables' heuristic in literature. Therefore, it can be used in the case of larger data sets and does not require huge amount of memory. Experimental results of depth, average depth and number of nodes of decision trees constructed by this algorithm are compared in the framework of each of the three approaches. 'Misclassification Error' Greedy Heuristic to Construct Decision Trees for Inconsistent Decision Tables.

We used decision tree as a model to discover the knowledge from multi-label decision tables where each row has a set of decisions attached to it and our goal is to find out one arbitrary decision from the set of decisions attached to a row. The size of the decision tree can be small as well as very large. We study here different greedy as well as dynamic programming algorithms to minimize the size of the decision trees. When we compare the optimal result from dynamic programming algorithm, we found some greedy algorithms produce results which are close to the optimal result for the minimization of number of nodes (at most 18.92% difference), number of nonterminal nodes (at most 20.76% difference), and number of terminal nodes (at most 18.71% difference).

Decision tree is a widely used technique to discover patterns from consistent data set. But if the data set is inconsistent, where there are groups of examples (objects) with equal values of conditional attributes but different decisions (values of the decision attribute), then to discover the essential patterns or knowledge from the data set is challenging. We consider three approaches (generalized, most common and many-valued decision) to handle such inconsistency. We created different greedy algorithms using various types of impurity and uncertainty measures to construct decision trees. We compared the three approaches based on the decision tree properties of the depth, average depth and number of nodes. Based on the result of the comparison, we choose to work with the many-valued decision approach. Now to determine which greedy algorithms are efficient, we compared them based on the optimization and classification results. It was found that some greedy algorithms (Mult_ws_entSort, and Mult_ws_entML) are good for both optimization and classification.

Wireless sensor network (WSN) is a wireless ad hoc network that consists of very large number of tiny sensor nodes communicating with each other with limited power and memory constrain. WSN demands real-time forwarding which means messages in the network are delivered according to their end-to-end deadlines (packet lifetime). This paper proposes a novel real-time routing protocol with load distribution (RTLD) that ensures high packet throughput with minimized packet overhead and prolongs the lifetime of WSN. The routing depends on optimal forwarding decision that takes into account of the link quality, packet delay time and the remaining power of next hop sensor nodes. The proposed mechanism has been successfully studied and verified through simulation and real test bed implementation.

The recognition-primed decision (RPD) model (Klein, 1993) is an account of expert decision making that focuses on how experts recognize situations as being similar to past experienced events and thus rely on memory and experience to make decisions. A number of computational models exist that attempt to account for similar aspects of expert decision making. In this paper, I briefly review these extant models and propose the Bayesian recognitional decision model (BRDM), a Bayesian implementation of the RPD model based primarily on models of episodic recognition memory (Mueller & Shiffrin, 2006; Shiffrin & Steyvers, 1997). The proposed model accounts for three important factors used by experts to make decisions: evidence about a current situation, the prior base rate of events in the environment, and the reliability of the information reporter. The Bayesian framework integrates these three aspects of information together in an optimal way and provides a principled framework for underst...

Decision rules provide a flexible toolbox for solving computationally demanding, multistage adaptive optimization problems. There is a plethora of realvalued decision rules that are highly scalable and achieve good quality solutions. On the other hand, existing binary decision rule structures tend to produce good quality solutions at the expense of limited scalability and are typically confined to worst-case optimization problems. To address these issues, we first propose a linearly parameterised binary decision rule structure and derive the exact reformulation of the decision rule problem. In the cases where the resulting optimization problem grows exponentially with respect to the problem data, we provide a systematic methodology that trades-off scalability and optimality, resulting to practical binary decision rules. We also apply the proposed binary decision rules to the class of problems with random-recourse and show that they share similar complexity as the fixed-recourse problems. Our numerical results demonstrate the effectiveness of the proposed binary decision rules and show that they are (i) highly scalable and (ii) provide high quality solutions.

It is now u n d e r s t o o d [2, 3 1 t h a t the t u r b o decoding algorithm is a n instance of a probability propagation a l g o r i t h m ( P P A ) o n a g r a p h w i t h m a n y cycles. I n this paper w e investigate t h e behavior of a n P P A in graphs w i t h a single cycle such as the g r a p h of a tail-biting code. First, w e show t h a t for strictly positive local kernels, t h e iterations of the P P A converge to a unique fixed point, (which was also observed b y Anderson a n d Hladik [l] a n d Weiss [ 5 ] ) . Secondly, w e shall generalize a result of McEliece a n d Rodemich [4], b y showing that if the hidden variables in t h e cycle are binary-valued, t h e P P A will always make a n o p t i m a l decision. (This was also observed independently by Weiss [ 5 ] ) . W h e n t h e hidden variables c a n a s s u m e 3 or m o r e values, t h e behavior of t h e P P A is much h a r d e r to characterize.

We describe the use of a successful combination of Bayesian inference and decision theory in a clinical trial design. The trial involves three important decisions, adaptive dose allocation, optimal stopping of the trial, and the optimal terminal decision upon stopping. For all three decisions we use a formal Bayesian decision-theoretic approach. The application demonstrates how Bayesian posterior inference and decision-theoretic approaches combine to provide a coherent solution in a complex application. The main challenges are the need for a flexible probability model for the unknown dose-response curve, a delayed response, the sequential nature of the stopping decision, and the complex considerations involved in the terminal decision. The main methodological features of the proposed solution are the use of decision theory to achieve optimal learning about the unknown dose-response curve, an innovative grid-based approximation method to implement backward induction for the sequential stopping decision, and a utility function for the terminal decision that is based on a posterior predictive description of a future confirmatory trial.

The concept of dependence among variables in a Bayesian belief network is well understood, but what does it mean in an influence diagram where some of those variables are decisions? There are three quite different answers to this question that take the familiar concepts for uncertain variables and extend them to decisions. First is responsiveness, whether the choice for a decision affects another variable. Second is materiality, whether observing other variables before making a decision affects the choice for the decision and thus improves its quality. Third is the usual notion of dependence, assuming that all of the decisions are made optimally given the information available at the time of the decisions. There are some subtleties involved, but all three types of decision dependence can be quite useful for understanding a decision model.

We consider a problem where an uninformed principal makes a timing decision interacting with an informed but biased agent. Because time is irreversible, the direction of the bias crucially affects the agent's ability to credibly communicate information. When the agent favors late decision making, full information revelation often occurs. In this case, centralized decision making, where the principal retains authority and communicates with the agent, implements the optimal decision-making rule. When the agent favors early decision making, communication is partial, and the optimal decision-making rule is not implemented. Delegation adds value when the bias is for early decision making, but not for late decision making. (JEL D21, D23, D82, D83) Many decisions in organizations deal with the optimal timing of taking a certain action. Because information in organizations is dispersed, the decision maker needs to rely on the information of her better-informed subordinates who, however, may have conflicting preferences. Consider the following two examples of such settings: (i) in a typical hierarchical firm, top executives may be less informed than the product manager about the optimal timing of the launch of a new product. It would not be surprising for an empire-building product manager to be biased in favor of an earlier launch; (ii) the CEO of a multinational corporation is contemplating when to shut down a plant in a struggling economic region. While the local plant manager is better informed about the prospects of the plant, he may be biased toward a later shutdown due to personal costs of relocation. These examples share a common theme. An uninformed principal faces an optimal stopping-time problem-when to exercise a real option. An agent is better informed than the principal but is biased toward earlier or later exercise. In this paper, we study how organizations make timing decisions in such a setting. We examine the

A decision space is defined by the range of options at the decision maker's disposal. For each option there is a distribution of possible consequences. Each distribution is a function of the uncertainty of elements in the decision situation (how big is the fire) ...

For decades, behavioral scientists have used the matching law to quantify how animals distribute their choices between multiple options in response to reinforcement they receive. More recently, many reinforcement learning (RL) models have been developed to explain choice by integrating reward feedback over time. Despite reasonable success of RL models in capturing choice on a trial-by-trial basis, these models cannot capture variability in matching. To address this, we developed novel metrics based on information theory and applied them to choice data from dynamic learning tasks in mice and monkeys. We found that a single entropy-based metric can explain 50% and 41% of variance in matching in mice and monkeys, respectively. We then used limitations of existing RL models in capturing entropy-based metrics to construct a more accurate model of choice. Together, our novel entropy-based metrics provide a powerful, model-free tool to predict adaptive choice behavior and reveal underlying...

This paper focus on how Decision Risk Analysis and VOI have been used to improve corporate investment decision on oil field development at PT. X. Decision and risk analysis is an ideal tool for decision making in an environment of risk and uncertainty. The process enables a more holistic view to be made within the context of the overall objective to get clear understanding of the oilfield asset development. In the petroleum industry, the decision to develop new field must be made under conditions of uncertainty. The degree of uncertainty and the economic consequences that associated with such investment decisions can vary widely. Fail in performing good decision could lead to inefficient of a multi-billion dollar capital investment and will ultimately affect the financial performance of the firm. Although uncertainty can always be reduced by simply buying more information, there must be a point at which the cost of additional information exceeds its benefit. This paper examines the problem of determining the point at which the development decision should be made. A complex decision problem is composed into separate components, form of a decision tree, on which perceptions of options, uncertainties and values can be explicitly represented. The methodology is applied to a real green-field development in Riau-Sumatera.

Normally a decision support system is build to solve problem where multicriteria decisions are involved. The knowledge base is the vital part of the decision support containing the information or data that is used in decision-making process. This is the field where engineers and scientists have applied several intelligent techniques and heuristics to obtain optimal decisions from imprecise information. In this paper, we present a hybrid neurogenetic learning approach for the adaptation a Mamdani fuzzy inference system for the Tactical Air Combat Decision Support System (TACDSS). Some simulation results demonstrating the difference of the learning techniques and are also provided.

We explore the optimal selection problem where two decision makers are involved in the evaluation of the arriving offers. We develop three stopping rules to avoid conflictual situations where a decision makers agrees with a current offer and the other wishes to discard it. The three stopping rules are then implemented in C language and runned on a series of problems with various sizes. The experimental results show that the selection rules generate different behaviours of the decision makers. Finally, we compare our problem with the optimal selection problem including a single decision maker and show the loss of utility generated by the introduction of an additional decision maker.

To optimize the decision functions, we will start setting parameters of decision functions, and then define some criteria that allow them to assess them performance. Two methods for construction of decision functions are proposed to meet the best criteria of performance: one is based on an algorithm derived from the formulation of the problem, the other is based on an optimization by a genetic algorithm.

The quality of a selection decision is a function of the decision rule used and the data collected to support the decision. When physical measurements are the basis of the decision data, the measurement sampling scheme controls measurement uncertainty and influences decision quality. Faced with a fixed experimental budget for measurement collection over multiple attributes, a decision-maker must decide how to allocate this budget to best support the selection decision. We expand on previous work in this area of sample allocation in multiple attribute selection decisions to compare the quality of allocation rules derived under various estimation and decision rules. The decision rules considered in this work include the selection of the best (expected value) and multinomial selection with allocation rules developed based on maximum likelihood and Bayesian inference. We derive allocation rules for each of these cases and illustrate their performance through computational experiments.

Manufacturers today are increasingly adopting a dual channel to sell their products, i.e., the traditional retail channel and an online direct channel. Empirical studies have shown that service quality (we focus on the delivery lead time of the direct channel) even goes beyond product price as one of the major factors influencing consumer acceptance of the direct channel. Delivery lead time has significant effects on demand, profit, and pricing strategy. However, there is scant literature addressing the decision on the promised delivery lead time of a direct channel and its impact on the manufacturer's and retailer's pricing decisions. To fill this gap, we examine the optimal decisions of delivery lead time and prices in a centralized and a decentralized dual-channel supply chain using the two-stage optimization technique and Stackelberg game, and analyze the impacts of delivery lead time and customer acceptance of a direct channel on the manufacturer's and retailer's pricing behaviours. We analytically show that delivery lead time strongly influences the manufacturer's and the retailer's pricing strategies and profits. Our numerical studies reveal that the difference between the demand transfer ratios in the two channels with respect to delivery lead time and direct sale price, customer acceptance of the direct channel, and product type have great effects on the lead time and pricing decisions.

Purpose-The purpose of this paper is to explore the extent to which outsourcing can be regarded as a mode of increasing organization learning through the internalization of new routines. Design/methodology/approach-The paper features six case studies of firms that have outsourced parts, or all of their information technology (IT) activities. Findings-When a firm outsources an IT activity (that has been performed in-house), it needs to develop an ability to specify to its supplier(s), and articulate its IT requirements in explicit terms. A firm's effort to externalize knowledge that was internal to an external supplier implies that a great deal of relatively tacit knowledge has to be converted into explicit knowledge, so that suppliers are able to understand the firm's business specificity. In this very process of externalizing knowledge and interacting with suppliers and other market players, the firm develops new rules, routines and procedures relating to how to manage the outsourced IT activity, which over time will be internalized. Originality/value-The paper is of value in linking outsourcing and organizational learning.

The chapter presents a decision support system. The decision-making process is modeled by a multi-criteria optimization problem. The decision support method is an interactive decision-making process. The choice is made by solving the problem depending on the control parameters that define the aspirations of the decision makers for each criteria function, then it evaluates the obtained solution by accepting or rejecting it. In another case, the decision maker selects a new value and the problem is solved again for the new parameter. In this chapter, an example of a decision support system is presented.

The paper presents the method of a group decision making in a competitive environment. We deal with a group decision when the group of people with different preferences are to make one single decision. The group decision selection process is modeled with the use of multi-criteria optimization task. It is solved with the use of reference point method. This method is an interactive method in which every person specifies its requirements in the form of a reference point, expressing the desired values for its evaluation function. On the basis of the provided reference point, a scalar achievement function is built. Maximization of this function generates a solution of the multicriteria task. This solution is presented to every person for acceptance or as a basis for the modification of the reference point. The paper contains the example of application of the proposed method to support a group decision by three people with different preferences.

The paper presents the method of facilitating joint decision making in a competitive environment. We deal with a joint decision when the group of people with different preferences are to make one single decision. The joint decision selection process is modelled with the use of multi-criteria optimization task. It is solved with the use of reference point method. This method is an interactive method in which every person specifies its requirements in the form of a reference point, expressing the desired values for its evaluation function. On the basis of the provided reference point, a scalar achievement function is built. Maximization of this function generates a solution of the multi-criteria task. This solution is presented to every person for acceptance or as a basis for the modification of the reference point. The paper gives an example of applying the proposed decision making while competing by three persons of different preferences.